Logo

A pattern-based algorithmic autotuner for graph processing on GPUs.

View the Project on GitHub PAA-NCIC/GSWITCH

GSWITCH

GSWITCH is a pattern-based algorithmic autotuning system that dynamically switched to the suitable optimization variants with negligible overhead. Specifically, It is a CUDA library targeting the GPU-based graph applications, it supports both vertex-centric or edge-centric abstractions. By far, GSWITCH can automatically determine the suitable optimization variants in Direction (push, pull), data-structure (Bitmap, Sorted Queue, Unsorted Queue), Load-Balance (TWC, WM, CM, STRICT, 2D-partition), Stepping (Increase, Decrease, Remain), and Kernel Fusion (Standalone, Fused). The fast optimization transition of GSWITCH is based on a machine learning model trained from 600+ real graphs from the network repository. The model can be resued by new applications, or be retrained to adapt to new architectures. In addition, GSWITCH provides succinct programming interface which hides all low-level tuning details. Developers can implements their graph applications with high performance in just ~100 lines of code.

Why GSWTICH

As GPUs provide higher parallelism and memory bandwidth than traditional CPUs, GPUs become a promising hardware to accelerate graph algorithms. Many recent works have explored the potential of using GPUs for data-intensive graph processing. Although the primary optimizations of these works are diverse, we notice that most of them are trying to find a ‘one size fits all’ solution. This leads to the mismatch and complication issues:

Mismatch: Previous GPU-based graph frameworks may incur performance hits due to suboptimal strategies. Previous works accelerated graph primitives to run truly fast on some particular graphs or algorithms, however, their performance might fell dramatically when facing an unmatched situation. For example, Figure 2 shows that differents graph require different load-balance strategies. Figure 3 shows the performance loss if we only use push model in frontier expansion.

LB Figure 1: Best load-balance for different graph

loss Figure 2: performance loss

Complication: Priori knowledge is required for users to make favorable decisions, especially from a mass of choices. A bulk synchronous parallel (BSP)-style graph application achieves its best performance only if correct strategies are chosen in every super-step. Unfortunately, the number of these performance-crucial strategies is very large, or worse yet, various combinations of these strategies form a huge tuning space. Data analysts should not spend their labor on wrestling with the tedious and complex performance tuning. Offloading the decision-making to a fully auto-tuning runtime could be a better choice.

Dependency

Build Instruction

Clone GSWITCH code to local server and build GSWITCH with CMake.

$ git clone https://github.com/PAA-NCIC/GSWITCH.git
$ cd GSWITCH
$ mkdir build && cd build
$ cmake ../ && make -j8

Usage

Here are the basic useages of pre-integrated applications (BFS, CC, PR, SSSP, BC) in GSWITCH.

./EXE <graph_path> [options]
[-r, --src=<int>]      Choose a root vertex. (Default: ch
                       oose randomly).
[-v, --verbose]        Print verbose per iteration info. 
                       (Default: quiet mode)
[-V, --validation]     Process the CPU reference validati
                       on. (Defaule: no validation)
[-H, --with-header]    Input file has header (e.g. nverte
                       xs, nvertexs, nedges, Default: no 
                       header).
[-W, --with-weight]    Input file has weight.(Default: no
                        weight value)
[-i, --ignore-weight]  Ignore the graph weight.(Default: 
                       false)
[-d, --directed]       Graph is directed.(Default: undire
                       cted)
[-c, --configs=Push-Queue-CM-Fused]
                       Set debug strategies, use - to sep
                       arater them (Default: <empty>).
[-D, --device=<int>]   Choose GPU for testing (Default: 0
                       )
[-j, --json=<string>]  set the json path (Default: 0)

*Note: By using --configs, you can force the applications to run with the static strategies. (No dynamic transition).

APIs

To customize your own application. you should provide at most six small functions.

APIs Description
filter(int vidx, G g) required, stream all the vertices(or edges) and filter out active ones; then update their value.
WA emit(int vidx, E* e, G g) required, describe the message from one vertex to another.
comp(WA* vdata, WA msg, G g) required, describe how the message is processed in the target vertex.
compAtomic(WA* vdata, WA msg, G g) ditto, but an atomic version.
cond(int vidx, E* e, G g) optional, help to omit useless updates.
exit(int vidx, E* e, G g) optional, customed exit condition.

Example

Here is a sample codes of BFS for graph soc-orkut. For more details please visit ./application/bfs.cu

#include "gswitch.h"
using G = device_graph_t<CSR, Empty>;

struct BFS:Functor<VC,int,Empty,Empty>{
  __device__ Status filter(int vid, G g){
    int lvl = *wa_of(vid);
    if(lvl == g.get_level()) return Active;
    else if (lvl < 0) return Inactive;
    else return Fixed;
  }
  __device__ int emit(int vid, Empty *w, G g) {return g.get_level();}
  __device__ bool cond(int v, int newv, G g) {return *wa_of(v)==-1;}
  __device__ bool comp(int* v, int newv, G g) {*v=newv; return true;}
  __device__ bool compAtomic(int* v, int newv, G g) {*v=newv; return true;}
};

int main(){
  // load graph
  for(level=0;;level++){
    inspector.inspect(as, g, f, stats, fets, conf);
    if(as.finish(g, f, conf)) break;
    selector.select(stats, fets, conf);
    executor.filter(as, g, f, stats, fets, conf);
    g.update_level();
    executor.expand(as, g, f, stats, fets, conf);
  }
  // copy data back
}

run the with ./BFS soc-orkut.mtx --with-header --src=0 --device=0 --verbose:

run-bfs Figure 3: BFS example

path Figure 4: the decision path of BFS for graph orkut

Applications

Typical applications

Here we describe five typical graph applications in our framework to show how to translate real applications into our filter-expand framework: the Breadth-first search algorithm (BFS), the connected components algorithm (CC), the PageRank algorithm (PR), the single source shortest path algorithm (SSSP), and the betweenness centrality algorithm (BC). These five benchmarks can cover the majority of real-world graph applications.

Breadth-First Search. A BFS algorithm generates the breadth-first search tree of a graph from a source vertex root and compute the shortest jump hops from the root to all vertices reachable from it. Both of each vertex and each edge will be processed at most once. In BFS, any race condition between edges visiting a vertex is benign, thus we can perform pruning when we find a vertex that has been touched in the current iteration.

In our framework, The BFS algorithm is implemented in vertex-centric abstraction. We use a filter function to mark a vertex whose label equals to the BFS depth as active and emit the label+1 to the inactive vertices. Because the update process of BFS is idempotent, both of the comp and compAtomic function can do the update without atomic operations.

Connected Components. For undirected graphs, we say a group of vertices as a connected component when all vertices in the group can reach one another. Soman shows that the edge-centric implementation may be a better choice. In the edge-centric mode, each edge of the graph tries to assign its two end vertices with the same ID. A pointer-jumping is used to reduce the multi-level tree into star-like, which ensures that once the two end vertices of an edge have the same ID, they will remain the same in the following iteration. By repeating these two steps until no vertex changes, the algorithm terminates and outputs the result.

We implement the CC algorithm also as edge-centric in GSWITCH. At the beginning, all the edges are active, and we use a emit function to make the lower vertex writes its ID to the higher vertex in the expansion step. After expansion, we add an extra pointer-jumping phase to perform the root-finding procedure. At last, the filter function will mark the edge whose two end vertices have the same ID as fixed and mark the other edges as active. When all the edges become fixed, the whole algorithm terminates.

PageRank. Google first introduced the PageRank algorithm to calculate the relative importance of webpages. Given a graph G=<V,E>, a damping factor d (0<d<1), the PR value of a vertex v is calculated by:

\[PR[v] = \frac{1-d}{|V|} + d\sum_{u\in V^-(v)}{\frac{PR[u]}{dg^+(u)}}\]

At first, the PR values of all the vertices are set all the same and then use the equation above to compute until all of the differences of PR values between iterations drop to below a threshold value. GraphLab has introduced a Delta-PageRank algorithm to only send their changes (deltas) in PR values to their neighbors, thus not all the vertices are active during each iteration, which means less data movement.

In GSWITCH, we implement both of the traditional PageRank and the Delta-PageRank. In traditional PageRank, the emit function send the PR value of a vertex to its neighbors, and the new PR value is calculated in the filter function. In Delta-PageRank, the emit function only send the deltas in the last iteration, and the filter function will accumulate the sum of delta contributions from their neighbors. When the accumulated delta is large enough, the vertex is marked as active and updates its PR value. Both of the two algorithms have the same comp and compAtomic functions. They accumulate the messages sent from neighbors.

Single-Source Shortest Path. An SSSP problem takes a weighted graph and a rooted vertex as input, and compute the shortest path distance from the root vertex to each vertex reachable from it. For graphs with negative weight, Bellman-Ford algorithm can report the existence of a negative cycle. For graphs with non-negative weight, the distance of each vertex decrease monotonously, the delta-stepping algorithm is used to make a balance between work-efficiency and parallelism. As described in a recent work (Davidson2014Work), they used a Near-Far classification to preferentially update a group of vertices based on distance scoring heuristic.

We implement the sssp algorithm with our dynamic window optimization to overcome the drawback of the irregular workloads across iterations. For an edge e connected vertices of v and u, the emit function send the sum of the distance value of v and the weight of e. Then we compare the received message and the distance of u in the comp and compAtomic function. Note that in compAtomic, we use atomicMin to ensure atomicity. Finally, we use a filter function to choose the vertices whose distance fall in the dynamic window.

Betweenness Centrality. Centrality is widely used to indecate the importance of nodes in social graphs. The commonest formulation of the BC problem is in Brandes’s work. For a graph G=<V, E>, and a root vertex root. Let the nsp[v] be the number of the shortest paths from root to v. The BC value of the parent vertex $p$ of the $v$ in the BFS tree rooted as root can be calculated according to:

\[BC[p] = \sum_{v\in child(p)}{\frac{nsp[p]}{nsp[v]}(BC[v]+1)}\]

The algorithm is composed of two phases: a forward phase and a backward phase. The forward phase computes the number of the shortest paths from the $root$ to each vertex in BFS order. The backward phase computes the BC value using the equation above in a reversed BFS order.

In GSWITCH, We also implement the forward and backward phases. In the forward phase, we use a filter function similar with BFS to choose the active vertices and use the emit function to send the number of the shortest paths to inactive vertices. The comp and compAtomic functions are used to sum the numbers of the shortest paths. In the backward phase. A filter function will choose and update the vertices according to the level computed in the forward phase. A emit function will send BC value to its neighbors, while the comp and compAtomic function do the sum work.

Extended applications

Besides the above five application, we will constantly updated other applictions here:

Graph Coloring. Graph coloring partitions the vertices of a graph such that no two adjacent vertices share the same color. In most of cases, applications relying on this algorithm do not require the optimal coloring, such as Pannotia. Doing such coloring is among the first steps in many parallel graph algorithms. In the initialization step, each vertex is labeled with a random integer value. The algorithm then launches multiple iterations, each responsible for labeling one color. For each vertex, the algorithm compares its vertex value with that of its neighboring vertices. If the vertex value of a given node happens to be the largest (or smallest) among its neighbors it marks itself with the current iteration colors (one each for the largest and smallest in each set). The algorithm terminates when all vertices are colored.

We implement the GC algorithm as vertex-centric in GSWITCH. We use the filter function to filter out the uncolored vertices as our active set, then we color the vertice whose local maximum vertex id collected in the last iteration is smaller then it’s own vertex id. The emit function is used to send vertex id to each vertex’s neighbors. The comp and compAtomic compute the max vertex id of each vertex’s neighbors. This naive implementation has many optimization such as multi-hash and min-max, which can make the algorithm converges faster.

Performance

Here we show 100 cases for each application (BFS,CC,PR,SSSP,BC) compared with Gunrock on k40m,P100,and V100. (P.S. Note that some graph below have the same names with the graph in other well-known dataset such as SNAP, but THEY ARE DIFFERENT.)

K40m

DOBFS (both without tuning parameters)

Dataset Gunrock GSWITCH
cvxqp3.mtx 5.87296485901 1.6582
SOPF_FS_b39_c7.mtx 6.72578811646 1.77515
bundle1.mtx 5.12003898621 1.36597
scircuit.mtx 16.7441368103 8.01489
fe-ocean.mtx 29.777765274 12.519
socfb-Vanderbilt48.mtx 2.52294540405 0.858887
sme3Da.mtx 25.4499912262 5.25195
neos1.mtx 2.70509719849 0.912842
crystk02.mtx 70.0058898926 10.2581
3Dspectralwave2.mtx 5.04207611084 2.27588
3D_51448_3D.mtx 32.0639610291 4.72998
ibm_matrix_2.mtx 36.4141464233 4.84082
g7jac140sc.mtx 18.6948776245 2.54028
case39.mtx 9.09209251404 2.06079
socfb-UMass92.mtx 4.33206558228 0.902344
fp.mtx 5.09595870972 1.87793
neos2.mtx 2.60519981384 0.83667
c-73.mtx 5.66005706787 1.91406
frb50-23-1.mtx 0.723123550415 0.360107
TF17.mtx 13.6611461639 1.479
bas1lp.mtx 4.16803359985 0.969971
mri1.mtx 3.86190414429 1.73706
g7jac160sc.mtx 31.3901901245 3.23877
shar_te2-b2.mtx 12.5088691711 1.56421
socfb-Mississippi66.mtx 3.36003303528 0.858887
g7jac180.mtx 16.1709785461 2.86084
g7jac180sc.mtx 15.9480571747 2.87598
in-2004.mtx 18.1181430817 8.17505
bcsstk35.mtx 31.7490100861 8.92505
nemeth24.mtx 104.316947937 25.0598
g7jac200sc.mtx 12.3958587646 2.83691
turon_m.mtx 18.4290409088 8.60693
gas_sensor.mtx 17.7128314972 7.93408
ASIC_680ks.mtx 12.7611160278 3.99805
sgpf5y6.mtx 4.42910194397 2.02197
pds-70.mtx 4.63199615479 2.25073
TSOPF_FS_b162_c3.mtx 7.78794288635 1.67993
invextr1_new.mtx 25.9199142456 7.88403
rajat21.mtx 66.0350341797 7.09497
HFE18_96_in.mtx 2.601146698 1.13599
cage12.mtx 6.62398338318 3.09692
onf5_4-8x8-20.mtx 41.0130004883 2.5918
conf6_0-8x8-20.mtx 6.90817832947 2.52881
conf6_0-8x8-30.mtx 37.868976593 2.59497
conf6_0-8x8-80.mtx 8.60190391541 2.56421
frb59-26-3.mtx 1.97196006775 0.406006
ca-dblp-2012.mtx 23.7967967987 2.88403
nemsemm1.mtx 11.0490322113 3.70801
wave.mtx 13.2989883423 6.13721
144.mtx 7.17806816101 3.52515
pkustk05.mtx 33.2131385803 14.8279
helm2d03.mtx 66.8899993896 31.7368
TSOPF_FS_b162_c4.mtx 9.97090339661 1.91211
Maragal_7.mtx 4.57882881165 0.829834
pkustk07.mtx 6.46114349365 2.88281
TSOPF_FS_b300_c2.mtx 153.057098389 51.9648
socfb-UIllinois.mtx 2.41589546204 0.942871
Ga19As19H42.mtx 6.99901580811 2.89795
matrix-new_3.mtx 32.1168899536 3.60107
sc-nasasrb.mtx 59.9720458984 28.8
pct20stif.mtx 154.378173828 5.68921
ramage02.mtx 34.6400756836 3.79004
m_t1.mtx 25.3880023956 9.48096
mouse_gene.mtx 5.56111335754 1.49414
roadNet-PA.mtx 238.497024536 35.9468
TSOPF_FS_b39_c30.mtx 27.0621776581 4.07202
mc2depi.mtx 247.810836792 115.603
3dtube.mtx 65.5870437622 16.9741
TF18.mtx 5.56802749634 2.27197
av41092.mtx 4.41408157349 1.573
cont1_l.mtx 7.62796401978 1.72119
watson_2.mtx 7.61318206787 3.54785
web-Stanford.mtx 64.7459030151 29.74
offshore.mtx 13.2050514221 4.51709
ecology2.mtx 276.221038818 138.065
pkustk04.mtx 89.2648696899 15.6208
gupta2.mtx 40.7769699097 10.949
t3dh_e.mtx 29.1409492493 5.25488
TSOPF_FS_b300.mtx 10.1289749146 1.60498
Si87H76.mtx 12.4158859253 5.67993
ship_001.mtx 38.3911132812 17.387
c8_mat11.mtx 5.42902946472 0.848145
s3dkq4m2.mtx 57.1749191284 25.7554
TF19.mtx 9.84597206116 2.64185
pattern1.mtx 11.3041400909 1.03394
gupta3.mtx 9.68503952026 0.89502
LargeRegFile.mtx 229.742050171 5.1499
tp-6.mtx 19.4079875946 2.04297
sc-pwtk.mtx 107.097862244 45.554
soc-digg.mtx 5.20586967468 2.4541
hugetrace-00000.mtx 414.326202393 179.371
web-it-2004.mtx 13.1268501282 4.60498
frb100-40.mtx 12.2449398041 0.401123
hugetric-00000.mtx 446.645019531 212.517
adaptive.mtx 954.504943848 415.347
co-papers-citeseer.mtx 25.2740383148 6.47705
packing-500x100x100-b050.mtx 121.479034424 50.6799
socfb-A-anon.mtx 14.4929885864 5.31885
delaunay_n23.mtx 229.893920898 101.023
channel-500x100x100-b050.mtx 157.440185547 56.637
delaunay_n24.mtx 345.646148682 164.176

CC (Both are edge-centric)

Dataset Gunrock GSWITCH
ct20stif.mtx 2.77304649353 0.63208
cbuckle.mtx 1.98793411255 0.968994
cvxbqp1.mtx 2.50697135925 1.03979
ch7-8-b4.mtx 2.80904769897 1.14209
darcy003.mtx 1.67512893677 0.755615
ch7-9-b3.mtx 2.39992141724 1.12109
dawson5.mtx 2.69603729248 0.722168
ch7-9-b5.mtx 9.52792167664 1.46704
dbir1.mtx 1.47414207458 0.62207
ch8-8-b3.mtx 1.36804580688 0.646729
dbir2.mtx 1.76286697388 0.635986
ch8-8-b4.mtx 1.38902664185 0.657959
dblp-2010.mtx 1.62696838379 0.733154
dc1.mtx 4.0328502655 1.6709
heart1.mtx 2.78997421265 1.38892
germany_osm.mtx 5.25307655334 1.14502
goodwin.mtx 2.57897377014 0.677002
graham1.mtx 3.44300270081 1.34717
graphics.mtx 2.92897224426 0.953125
great-britain_osm.mtx 2.4049282074 1.14014
heart3.mtx 2.35795974731 1.15308
gupta1.mtx 2.17509269714 0.86499
gupta2.mtx 5.2318572998 2.11401
gupta3.mtx 3.25798988342 0.824219
hood.mtx 4.09007072449 1.32397
email-EuAll.mtx 3.44395637512 0.971924
gyro.mtx 4.87303733826 1.03076
fem_filter.mtx 3.52478027344 1.44897
halfb.mtx 2.84600257874 1.15796
lp_ken_18.mtx 1.88207626343 0.719727
kkt_power.mtx 5.60092926025 1.89819
kneser_10_4_1.mtx 3.20506095886 1.44092
g7jac200.mtx 2.00390815735 0.779053
laminar_duct3D.mtx 1.64699554443 0.73291
g7jac200sc.mtx 2.83908843994 0.811035
landmark.mtx 3.20196151733 1.26562
misc-IMDB-bi.mtx 3.06010246277 1.52905
mixtank_new.mtx 3.10087203979 1.45825
mk12-b4.mtx 3.42392921448 0.9021
pwtk.mtx 5.30791282654 1.48877
raefsky2.mtx 12.0389461517 5.05908
raefsky4.mtx 8.21399688721 3.27686
raefsky5.mtx 4.82296943665 1.76099
rail4284.mtx 4.67991828918 2.03809
rail_79841.mtx 12.1219158173 4.26196
rajat15.mtx 13.5071277618 5.70386
rajat18.mtx 6.10113143921 2.2019
rajat21.mtx 3.91912460327 1.93018
rajat22.mtx 6.81781768799 1.96606
rajat24.mtx 3.51905822754 1.17627
rajat25.mtx 3.85785102844 1.83301
rajat26.mtx 4.24194335938 2.05713
rajat29.mtx 3.48806381226 1.69604
rajat30.mtx 5.11598587036 1.64795
rajat31.mtx 3.94797325134 1.33203
ramage02.mtx 3.20506095886 1.31421
rel8.mtx 3.7841796875 1.31104
relat8.mtx 2.83789634705 1.20581
relat9.mtx 4.83107566833 2.33887
rim.mtx 7.58290290833 3.13306
roadNet-PA.mtx 4.99796867371 1.77295
roadNet-TX.mtx 3.13305854797 1.2561
s1rmq4m1.mtx 4.13298606873 1.3811
s2rmq4m1.mtx 3.08012962341 1.30396
s2rmt3m1.mtx 12.5458240509 1.44409
s3rmq4m1.mtx 4.00805473328 1.81982
s3rmt3m1.mtx 16.0698890686 3.58105
s4dkt3m2.mtx 15.3570175171 7.37305
scircuit.mtx 15.6872272491 7.22681
shallow_water1.mtx 4.41002845764 1.79883
shallow_water2.mtx 9.36603546143 2.34106
shar_te2-b3.mtx 4.53901290894 1.55615
ship_001.mtx 3.2639503479 1.48096
ship_003.mtx 5.72085380554 1.68018
shipsec1.mtx 6.16502761841 1.90625
shipsec8.mtx 6.31785392761 1.8772
shyy161.mtx 9.32383537292 3.2749
sinc12.mtx 5.74588775635 1.43604
sinc15.mtx 5.76305389404 2.42798
sinc18.mtx 3.32999229431 1.39185
sls.mtx 3.99708747864 1.46387
sme3Db.mtx 3.82280349731 1.47192
sme3Dc.mtx 3.29279899597 1.58179
soc-LiveJournal1.mtx 3.3700466156 1.4751
t0331-4l.mtx 16.6139602661 2.94287
soc-Slashdot0811.mtx 13.1039619446 4.91895
soc-Slashdot0902.mtx 24.491071701 6.62109
soc-sign-Slashdot090216.mtx 16.2551403046 7.62207
soc-sign-Slashdot090221.mtx 5.95283508301 2.33105
t3dh.mtx 3.51691246033 1.54297
turon_m.mtx 5.90896606445 2.58496
twotone.mtx 6.28089904785 2.56299
vanbody.mtx 8.50200653076 3.26782
venkat01.mtx 8.58092308044 3.37793
venkat25.mtx 8.06283950806 2.68311
vfem.mtx 45.9461212158 18.7581
socfb-Columbia2.mtx 10.1180076599 3.64478
viscorocks.mtx 26.5548229218 6.375
socfb-Cornell5.mtx 36.0109806061 17.2588
water_tank.mtx 18.089056015 6.35889
socfb-Duke14.mtx 63.472032547 19.2708

PageRank (with the same threshold)

Dataset Gunrock GSWITCH
3D_51448_3D.mtx 11.2380981445 4.72192
g7jac140.mtx 12.6740932465 5.61499
hamming10-2.mtx 10.3750228882 3.84302
Raj1.mtx 41.9518947601 8.49902
Maragal_6.mtx 13.1239891052 6.34595
lp1.mtx 183.020114899 32.4751
bcsstk37.mtx 11.9321346283 4.4729
patents_main.mtx 83.2149982452 8.35303
frb50-23-2.mtx 11.0521316528 3.91821
hvdc2.mtx 12.9191875458 6.06494
soc-LiveJournal1.mtx 635.165929794 19.156
stat96v1.mtx 216.727018356 7.70361
pds-50.mtx 103.451013565 7.41699
rail2586.mtx 239.165067673 7.85083
cage13.mtx 174.88694191 6.36206
tmt_unsym.mtx 30.3120613098 5.32202
g7jac160sc.mtx 13.3528709412 6.36816
shar_te2-b2.mtx 280.174016953 5.56519
tech-RL-caida.mtx 18.2960033417 8.03223
thermomech_dM.mtx 36.1630916595 5.78906
boyd2.mtx 72.9038715363 24.9932
nemeth22.mtx 14.3570899963 4.4812
soc-twitter-follows.mtx 113.15202713 16.3601
TSOPF_RS_b2052_c1.mtx 13.6208534241 5.01294
pds-60.mtx 127.330064774 9.78198
rgg_n_2_17_s0.mtx 15.4240131378 6.97412
torso3.mtx 101.810932159 6.40479
nemeth23.mtx 11.8088722229 4.698
ASIC_320ks.mtx 59.8199367523 11.7651
Lin.mtx 36.6859436035 5.61621
NotreDame_www.mtx 143.15199852 10.0251
delaunay_n18.mtx 43.7529087067 6.88916
language.mtx 176.179885864 12.626
qa8fm.mtx 14.3330097198 6.28003
rajat24.mtx 103.203058243 15.6089
shar_te2-b3.mtx 14.6470069885 7.27686
dblp-2010.mtx 108.412981033 11.563
mk13-b5.mtx 63.9169216156 6.79297
ca-MathSciNet.mtx 210.743188858 12.063
ASIC_680ks.mtx 77.1999359131 12.1309
socfb-Harvard1.mtx 15.7029628754 7.302
venkat50.mtx 13.2749080658 5.90308
ns3Da.mtx 14.6100521088 7.1272
sgpf5y6.mtx 84.9740505219 7.58691
pds-70.mtx 144.984006882 12.1709
mono_500Hz.mtx 16.0708427429 7.51318
cfd1.mtx 14.0058994293 6.60376
rajat29.mtx 166.009902954 23.2021
rajat21.mtx 115.41891098 20.0891
pds-80.mtx 161.581993103 13.7842
darcy003.mtx 65.4871463776 9.15601
mario002.mtx 68.2969093323 9.23975
cage12.mtx 18.6719894409 8.12598
coAuthorsDBLP.mtx 45.9468364716 12.2029
mixtank_new.mtx 13.867855072 6.58618
kneser_10_4_1.mtx 68.2787895203 11.219
c-big.mtx 276.998996735 11.4849
atmosmodd.mtx 49.9241352081 7.31396
atmosmodj.mtx 57.893037796 7.31396
pds-90.mtx 177.029132843 14.313
conf6_0-8x8-80.mtx 14.4059658051 6.8999
neos.mtx 580.847024918 15.8191
frb59-26-2.mtx 18.6479091644 5.78711
watson_1.mtx 106.513023376 9.35425
dbic1.mtx 22.0100879669 10.8318
web-NotreDame.mtx 25.908946991 12.437
pds-100.mtx 190.760850906 15.6812
Freescale1.mtx 554.043054581 15.637
socfb-MSU24.mtx 22.9361057281 9.96118
connectus.mtx 503.82900238 33.916
dbir2.mtx 20.4219818115 9.23389
helm2d03.mtx 59.7171783447 10.2249
thermal2.mtx 94.3360328674 11.5869
soc-delicious.mtx 479.265928268 19.262
flickr.mtx 320.264101028 22.3999
circuit5M.mtx 662.358999252 22.2041
nlpkkt120.mtx 543.653011322 10.186
Si41Ge41H72.mtx 22.9659080505 11.3999
pkustk03.mtx 17.9071426392 8.92993
inf-roadNet-PA.mtx 101.7100811 15.4719
roadNet-PA.mtx 113.131999969 16.2612
ljournal-2008.mtx 1490.26417732 24.2891
inf-belgium_osm.mtx 117.704153061 19.1069
belgium_osm.mtx 109.932899475 19.188
delaunay_n19.mtx 83.4739208221 12.1411
mc2depi.mtx 72.3390579224 10.1758
parabolic_fem.mtx 88.7501239777 11.2102
Hamrle3.mtx 136.647939682 12.6802
Rucci1.mtx 780.635118484 16.3708
great-britain_osm.mtx 132.483005524 22.6492
cont1_l.mtx 702.295064926 130.175
GL7d17.mtx 621.984004974 26.2939
t2em.mtx 89.1511440277 13.2307
watson_2.mtx 229.428052902 18.553
roadNet-TX.mtx 132.482051849 19.7939
GL7d22.mtx 442.041873932 20.688
GL7d16.mtx 628.221988678 25.3667
ecology2.mtx 101.673126221 13.885
ecology1.mtx 98.906993866 13.8862
webbase-1M.mtx 560.477018356 30.0222
apache2.mtx 93.6460494995 13.0732

SSSP (Both enable stepping)

Dataset Gunrock GSWITCH
para-6.mtx 1.99699401855 0.929932
para-9.mtx 2.05898284912 1.01611
hvdc1.mtx 7.55000114441 3.28394
bauru5727.mtx 15.6869888306 4.87988
cavity17.mtx 45.8748321533 7.22021
cavity18.mtx 14.4829750061 5.86914
cavity20.mtx 13.1051540375 5.89819
cavity22.mtx 13.090133667 5.979
cavity24.mtx 14.9431228638 6.05029
cavity26.mtx 14.4040584564 5.62134
Kemelmacher.mtx 4.13799285889 2.06104
graphics.mtx 15.5980587006 2.51416
c-61.mtx 5.62405586243 2.32593
s1rmq4m1.mtx 15.4628753662 7.125
coater2.mtx 18.4071063995 7.71191
fem_hifreq_circuit.mtx 28.4621715546 13.9224
c-56.mtx 12.2361183167 4.72998
ncvxqp5.mtx 8.85891914368 4.23413
helm3d01.mtx 13.8421058655 4.64893
onetone2.mtx 9.84501838684 4.61597
graham1.mtx 12.6340389252 5.04199
inlet.mtx 134.886032104 27.9131
ncvxqp3.mtx 8.66389274597 4.15625
ex40.mtx 35.7789993286 9.49316
c-67.mtx 18.7311172485 6.06299
deltaX.mtx 20.9989547729 3.05615
c-68.mtx 21.7459201813 7.40894
epb3.mtx 100.191833496 48.063
fxm4_6.mtx 8.71682167053 4.24194
FEM_3D_thermal2.mtx 74.4819641113 33.832
mark3jac120.mtx 20.9710597992 7.92773
mark3jac120sc.mtx 20.5562114716 7.9209
c-69.mtx 15.9959793091 7.14575
c-70.mtx 28.1360149384 5.41309
c-72.mtx 14.356136322 6.28101
mark3jac140sc.mtx 21.8830108643 8.83105
email-EuAll.mtx 18.7590122223 5.49683
heart2.mtx 13.9570236206 3.14014
image_interp.mtx 93.9931869507 46.0579
RFdevice.mtx 8.47601890564 3.99292
dc2.mtx 18.0118083954 3.23193
flower_8_4.mtx 8.55183601379 4.20117
c-71.mtx 19.1378593445 8.46826
nemeth19.mtx 87.0599746704 41.218
fe_ocean.mtx 52.4818878174 23.8398
cont-300.mtx 134.707931519 59.4211
fe-tooth.mtx 23.0369567871 9.19873
fe_tooth.mtx 22.8610038757 9.2251
2D_54019_highK.mtx 34.0430755615 13.021
Dubcova2.mtx 38.459777832 15.2349
3Dspectralwave2.mtx 25.2449512482 5.37671
case39.mtx 16.0081386566 4.35571
c-73.mtx 22.5808620453 5.95605
frb50-23-1.mtx 7.50207901001 2.1499
bas1lp.mtx 20.1640129089 5.13013
rail2586.mtx 23.0309963226 8.84619
cage13.mtx 14.0771865845 6.42993
boyd2.mtx 34.9929351807 17.1699
li.mtx 20.7870006561 10.0059
TSOPF_RS_b39_c19.mtx 20.1978683472 4.65405
vfem.mtx 31.6431522369 13.033
soc-twitter-follows.mtx 49.0310211182 7.83105
rgg_n_2_17_s0.mtx 138.471130371 65.7761
598a.mtx 22.7079391479 10.314
d_pretok.mtx 97.1269607544 24.9761
turon_m.mtx 51.9452095032 21.1318
TSOPF_RS_b162_c4.mtx 11.5170478821 4.1748
CO.mtx 14.4340991974 6.22021
language.mtx 28.7320613861 13.5547
venkat01.mtx 29.2029380798 13.386
venkat25.mtx 28.9130210876 13.729
venkat50.mtx 28.5458564758 12.2239
cfd1.mtx 35.8240585327 17.271
appu.mtx 11.3160610199 4.98608
darcy003.mtx 283.935058594 103.494
mario002.mtx 232.207061768 104.589
cage12.mtx 20.2748775482 9.62085
net100.mtx 13.2060050964 6.22876
TSC_OPF_1047.mtx 25.377035141 9.29712
af_shell2.mtx 63.159942627 31.4221
atmosmodj.mtx 81.4990997314 36.4758
water_tank.mtx 37.1270179749 14.741
conf5_4-8x8-05.mtx 45.0170059204 4.90186
conf5_4-8x8-10.mtx 38.7840270996 4.85083
conf5_4-8x8-15.mtx 15.22397995 4.8501
conf5_4-8x8-20.mtx 15.4058933258 4.8667
conf6_0-8x8-20.mtx 19.6619033813 4.89722
conf6_0-8x8-30.mtx 57.4040412903 4.82007
conf6_0-8x8-80.mtx 12.1150016785 4.84912
H2O.mtx 18.8748836517 7.59497
Hook_1498.mtx 29.0341377258 11.9731
connectus.mtx 14.9919986725 5.77002
crashbasis.mtx 353.230010986 42.5957
majorbasis.mtx 106.11390686 42.4351
helm2d03.mtx 166.405914307 56.8552
mac_econ_fwd500.mtx 166.723007202 42.0881
cop20k_A.mtx 57.6269607544 17.2839
filter3D.mtx 67.2008972168 33.5322
ct20stif.mtx 25.7580280304 9.23901
pct20stif.mtx 21.7549800873 9.0769
socfb-Penn94.mtx 12.9871368408 6.48999

BC

Dataset Gunrock GSWITCH
g7jac040.mtx 18.385887146 3.14819
g7jac040sc.mtx 18.0320739746 3.09229
graphics.mtx 3.72290611267 1.65112
shallow_water2.mtx 94.0580368042 45.4141
chipcool1.mtx 15.0260925293 7.50708
circuit_4.mtx 9.19985771179 4.55908
c-61.mtx 6.90698623657 2.52686
raefsky1.mtx 9.58585739136 4.63477
poisson3Da.mtx 16.6900157928 7.98193
garon2.mtx 24.6660709381 11.2969
helm3d01.mtx 11.8319988251 4.46387
c-59.mtx 21.595954895 3.54321
lhr10c.mtx 35.4959945679 3.38306
lhr11c.mtx 8.07499885559 3.45264
c-67.mtx 5.12003898621 2.29785
c-67b.mtx 5.1441192627 2.271
deltaX.mtx 4.64200973511 2.04517
c-62.mtx 14.5308971405 2.21826
nd6k.mtx 11.3220214844 5.41382
ncvxqp7.mtx 12.7639770508 4.04565
lung2.mtx 113.388061523 54.4592
bayer01.mtx 5.33008575439 2.61475
sinc12.mtx 13.4980678558 3.07568
EAT_SR.mtx 77.9559631348 2.26709
lhr14c.mtx 30.8728218079 4.05615
mk12-b4.mtx 4.11581993103 1.93848
n4c6-b12.mtx 4.47702407837 2.10889
heart2.mtx 5.54609298706 2.70483
pds-30.mtx 5.62596321106 2.79321
g7jac100sc.mtx 33.3449859619 4.36548
socfb-Cal65.mtx 17.3230171204 2.83228
lp_ken_18.mtx 8.3920955658 3.24927
soc-slashdot.mtx 8.59308242798 4.16309
socfb-Bingham82.mtx 6.3648223877 2.8877
RFdevice.mtx 5.85889816284 2.34131
dc2.mtx 20.0400352478 1.92773
dc1.mtx 4.13012504578 1.94727
flower_8_4.mtx 5.85007667542 2.74097
lhr17.mtx 25.7549285889 3.50903
psmigr_2.mtx 4.82797622681 1.66772
psmigr_3.mtx 15.5189037323 2.08545
cit-HepPh.mtx 12.1970176697 4.4082
socfb-Vanderbilt48.mtx 6.68096542358 2.67383
socfb-UCF52.mtx 6.24513626099 2.698
g7jac120.mtx 25.1448154449 5.43384
stormg2-125.mtx 7.38501548767 3.41333
fome21.mtx 8.3920955658 3.73193
socfb-GWU54.mtx 15.0549411774 3.22876
scc_twitter-copen.mtx 8.60500335693 2.59009
socfb-JMU79.mtx 12.5648975372 3.22998
socfb-Northwestern25.mtx 13.1771564484 3.24756
socfb-Duke14.mtx 14.0700340271 3.49585
case39.mtx 12.2880935669 3.68408
socfb-UMass92.mtx 18.9130306244 2.90063
socfb-UC33.mtx 13.1318569183 5.91504
fp.mtx 9.06300544739 4.48364
socfb-NotreDame57.mtx 6.90412521362 3.19897
boyd1.mtx 7.52687454224 3.11206
frb50-23-2.mtx 5.78784942627 1.42261
frb50-23-5.mtx 3.18813323975 1.46704
TF17.mtx 8.5551738739 3.17822
stat96v1.mtx 28.0420780182 3.61108
g7jac160.mtx 29.0489196777 5.70093
shar_te2-b2.mtx 17.4961090088 7.35889
socfb-UConn.mtx 28.5861492157 3.41406
socfb-UConn91.mtx 11.9259357452 3.38623
tech-RL-caida.mtx 18.9678668976 5.9502
socfb-Mississippi66.mtx 6.03199005127 2.75195
socfb-BU10.mtx 18.3501243591 3.23022
SiO.mtx 18.1999206543 4.69629
socfb-MU78.mtx 32.1409683228 2.87305
Trec13.mtx 4.96912002563 2.44385
socfb-Baylor93.mtx 9.31692123413 3.28613
bibd_17_8.mtx 5.17201423645 1.77124
socfb-UPenn7.mtx 12.5470161438 3.36426
socfb-Virginia63.mtx 13.099193573 3.31812
ch7-8-b4.mtx 7.75694847107 3.16602
frb53-24-4.mtx 3.2069683075 1.40503
frb53-24-2.mtx 3.32713127136 1.39526
frb53-24-3.mtx 6.68001174927 1.37012
socfb-NYU9.mtx 17.2910690308 3.24414
n4c6-b6.mtx 6.61706924438 3.03979
socfb-Maryland58.mtx 28.785943985 2.65503
socfb-UCLA.mtx 5.82218170166 2.84351
socfb-UCLA26.mtx 6.05010986328 2.89697
g7jac200sc.mtx 14.6560668945 6.68311
socfb-Tennessee95.mtx 8.18705558777 3.05298
m133-b3.mtx 18.2960033417 3.93774
ca-MathSciNet.mtx 17.8940296173 6.7312
rel8.mtx 26.0708332062 4.00903
socfb-Harvard1.mtx 15.2611732483 3.58618
fem_filter.mtx 162.341125488 24.7131
n4c6-b11.mtx 7.31992721558 3.60474
socfb-Wisconsin87.mtx 14.2209529877 3.97876
nw14.mtx 7.80200958252 2.82007
socfb-Auburn71.mtx 8.47315788269 3.10718
C2000-5.mtx 14.2478942871 1.75098
conf5_4-8x8-10.mtx 30.2400588989 7.21802
conf6_0-8x8-80.mtx 18.1341171265 6.35425
socfb-FSU53.mtx 19.3450450897 3.05591
net4-1.mtx 61.3079071045 12.9727

P100

DOBFS (both without tuning parameters)

Dataset Gunrock GSWITCH
nemeth20.mtx 26.652097702 10.20605454
Dubcova2.mtx 11.6810798645 4.36328122
3Dspectralwave2.mtx 2.97403335571 1.45361418
olafu.mtx 19.8850631714 6.82104474
gyro.mtx 10.0150108337 4.2441405
gyro_k.mtx 10.0929737091 4.41406198
g7jac140.mtx 4.03499603271 0.87695314
hamming10-2.mtx 0.43511390686 0.1887207
socfb-UMass92.mtx 0.907897949219 0.43798824
Raj1.mtx 14.1229629517 4.78686518
socfb-UC33.mtx 0.942945480347 0.4628905
MANN-a45.mtx 0.486850738525 0.1638184
neos2.mtx 1.73497200012 0.35400394
lp1.mtx 1.45602226257 0.3872071
c-73.mtx 2.16698646545 0.90722658
c-73b.mtx 2.20084190369 0.90307584
patents_main.mtx 4.5280456543 1.8073733
bcsstk37.mtx 11.9569301605 4.37866214
bcsstk36.mtx 10.9729766846 2.973144
viscorocks.mtx 31.6638946533 10.3520516
msc23052.mtx 10.9031200409 2.65893544
p-hat1500-2.mtx 0.513076782227 0.21215824
mri2.mtx 13.9870643616 5.8405768
frb50-23-4.mtx 0.477075576782 0.20166018
frb50-23-1.mtx 0.490188598633 0.20532228
frb50-23-5.mtx 0.492095947266 0.2241212
nemeth21.mtx 22.5188732147 10.5688482
TF17.mtx 3.59296798706 0.94311504
bas1lp.mtx 1.05500221252 0.426514
IG5-16.mtx 0.960111618042 0.448242
mri1.mtx 2.46596336365 1.01049834
pds-50.mtx 2.20513343811 0.8691406
rail2586.mtx 8.1901550293 2.4809578
g7jac160sc.mtx 4.1880607605 0.9565432
lp_osa_30.mtx 1.78503990173 0.42919864
shar_te2-b2.mtx 1.96003913879 0.7041019
socfb-UConn.mtx 0.952005386353 0.4631348
tech-caidaRouterLevel.mtx 2.60901451111 1.29199274
caidaRouterLevel.mtx 2.56419181824 1.24902284
msc10848.mtx 5.23495674133 1.73706084
li.mtx 8.21208953857 3.01367124
pli.mtx 7.82990455627 2.69628994
laminar_duct3D.mtx 14.2209529877 4.9677731
Trec13.mtx 1.01208686829 0.40698248
raefsky4.mtx 9.95993614197 3.7761226
fe_rotor.mtx 8.99791717529 3.41845718
vfem.mtx 10.8270645142 4.41015594
rim.mtx 18.4030532837 9.177978
socfb-Baylor93.mtx 0.976800918579 0.3762207
bibd_17_8.mtx 0.489950180054 0.20678714
g7jac180.mtx 4.63390350342 1.0327148
g7jac180sc.mtx 4.55522537231 0.965332
in-2004.mtx 10.6711387634 3.6787096
ch7-8-b4.mtx 2.6068687439 0.58666988
bcsstk35.mtx 13.1001472473 4.05981474
soc-twitter-follows.mtx 1.6610622406 0.8010259
crankseg_1.mtx 7.14087486267 2.8835439
frb53-24-4.mtx 0.426054000854 0.1958007
frb53-24-2.mtx 0.481128692627 0.21435544
frb53-24-1.mtx 0.531911849976 0.20727538
frb53-24-5.mtx 0.478982925415 0.2072754
frb53-24-3.mtx 0.494956970215 0.2268067
ca-dblp-2010.mtx 3.33309173584 1.4064942
dblp-2010.mtx 3.66592407227 1.42187438
pds-60.mtx 2.33292579651 0.84619144
raefsky3.mtx 15.1948928833 5.1237814
598a.mtx 6.33096694946 2.48291028
NotreDame_www.mtx 6.96992874146 2.270019
socfb-UCLA.mtx 0.808954238892 0.3869628
Lin.mtx 31.7990779877 6.70434514
d_pretok.mtx 22.4421024323 6.5578615
lhr34.mtx 4.83298301697 1.42651344
lhr34c.mtx 4.85777854919 1.4936521
g7jac200.mtx 4.94599342346 1.03442378
g7jac200sc.mtx 4.79102134705 1.08984388
2cubes_sphere.mtx 6.5929889679 2.49414034
pkustk09.mtx 13.3278369904 5.2863763
turon_m.mtx 18.5060501099 6.69897574
delaunay_n18.mtx 30.613899231 6.20336828
socfb-UVA16.mtx 1.23381614685 0.48022454
language.mtx 4.4379234314 1.68676784
qa8fk.mtx 14.0810012817 5.71874984
qa8fm.mtx 13.8339996338 5.91626
rajat24.mtx 12.1910572052 2.87646468
shar_te2-b3.mtx 2.11691856384 0.68139682
m133-b3.mtx 2.06112861633 0.67138684
mk13-b5.mtx 1.78718566895 0.7424318
gas_sensor.mtx 13.8509273529 5.74414044
ASIC_680ks.mtx 8.25190544128 3.318848
socfb-Harvard1.mtx 1.1088848114 0.5205079
venkat01.mtx 12.8231048584 5.72168024
venkat25.mtx 12.4588012695 5.62475594
venkat50.mtx 12.7358436584 5.78247074
fem_filter.mtx 7.6858997345 3.13696254
ns3Da.mtx 6.07895851135 1.88037044
sgpf5y6.mtx 2.58994102478 0.95898458
pds-70.mtx 2.33101844788 0.9187012
mono_500Hz.mtx 12.3541355133 5.0187991
ch7-8-b5.mtx 2.66408920288 0.59912098
p-hat1500-3.mtx 0.43797492981 0.1926269
socfb-Berkeley13.mtx 0.963926315308 0.40380864

CC (Both are edge-centric)

Dataset Gunrock GSWITCH
lung2.mtx 2.06685066223 0.61792
nemeth14.mtx 1.56402587891 0.723145
m133-b3.mtx 1.78098678589 0.639893
olesnik0.mtx 1.54590606689 0.568115
mac_econ_fwd500.mtx 1.12390518188 0.348145
nemeth15.mtx 1.59478187561 0.64502
majorbasis.mtx 1.88302993774 0.447021
nemeth16.mtx 1.83391571045 0.598877
nemeth17.mtx 1.797914505 0.676025
mark3jac040.mtx 5.79118728638 1.84106
nemeth02.mtx 1.98006629944 0.602051
mark3jac040sc.mtx 0.811100006104 0.39917
nemeth18.mtx 3.73697280884 0.530029
mark3jac060.mtx 1.56402587891 0.603027
nemeth03.mtx 1.14989280701 0.385986
mark3jac060sc.mtx 1.58882141113 0.561035
nemeth19.mtx 0.897884368896 0.360107
mark3jac080.mtx 1.31297111511 0.460205
nemeth04.mtx 1.3267993927 0.501221
mark3jac080sc.mtx 2.50196456909 0.709961
nemeth20.mtx 3.92293930054 0.770264
mark3jac100.mtx 1.6131401062 0.617188
mark3jac100sc.mtx 1.33299827576 0.594238
nemeth21.mtx 1.99389457703 0.578857
mark3jac120.mtx 3.64899635315 0.953125
soc-sign-Slashdot081106.mtx 2.09212303162 0.614258
mark3jac120sc.mtx 3.62515449524 0.644043
nemeth06.mtx 1.37996673584 0.37793
mark3jac140sc.mtx 1.2378692627 0.362793
nemeth22.mtx 1.8138885498 0.547852
matrix-new_3.mtx 1.54304504395 0.704102
nemeth23.mtx 1.19304656982 0.371094
matrix_9.mtx 3.66306304932 0.925049
nemeth24.mtx 2.22706794739 0.626709
mc2depi.mtx 6.41202926636 0.626221
nemeth25.mtx 1.61910057068 0.619873
nemeth26.mtx 2.26306915283 0.578857
mesh_deform.mtx 1.88803672791 0.601074
onetone1.mtx 1.62792205811 0.577881
mip1.mtx 2.47716903687 0.593262
nemsemm1.mtx 2.26879119873 0.875977
misc-IMDB-bi.mtx 1.22618675232 0.586914
neos.mtx 1.39021873474 0.451904
mixtank_new.mtx 2.50792503357 0.420898
neos1.mtx 2.05183029175 0.624268
mk12-b4.mtx 2.12001800537 0.618164
mk13-b5.mtx 1.26218795776 0.417725
neos3.mtx 1.30605697632 0.410156
mono_500Hz.mtx 2.82597541809 1.19678
net100.mtx 2.68721580505 0.633057
net125.mtx 1.44505500793 0.605957
net150.mtx 1.3530254364 0.448975
onetone2.mtx 2.06589698792 0.654053
mri1.mtx 2.31289863586 0.614014
opt1.mtx 3.11398506165 1.31201
mri2.mtx 1.59788131714 0.657959
net4-1.mtx 3.25608253479 1.08105
msc10848.mtx 5.47194480896 1.5769
net75.mtx 3.6518573761 0.719971
msc23052.mtx 1.20401382446 0.403076
netherlands_osm.mtx 1.38592720032 0.422119
msdoor.mtx 1.44696235657 0.643066
mult_dcop_02.mtx 2.23398208618 0.418945
mult_dcop_03.mtx 2.25186347961 0.655029
ncvxqp5.mtx 1.83987617493 0.736816
n4c6-b10.mtx 2.04491615295 0.658936
nlpkkt80.mtx 2.31194496155 0.635986
n4c6-b11.mtx 2.27284431458 0.852783
nmos3.mtx 3.92699241638 0.637939
n4c6-b12.mtx 4.11701202393 1.729
ns3Da.mtx 4.42910194397 1.73901
n4c6-b6.mtx 4.83512878418 2.26489
nsct.mtx 1.91116333008 0.61792
n4c6-b7.mtx 2.35080718994 0.810791
nw14.mtx 4.41884994507 1.99609
n4c6-b8.mtx 4.45103645325 1.97803
offshore.mtx 1.85203552246 0.705811
ohne2.mtx 4.11200523376 1.09912
nasasrb.mtx 1.43599510193 0.658936
oilpan.mtx 4.6079158783 1.10303
ncvxbqp1.mtx 0.990867614746 0.446045
olafu.mtx 2.84600257874 0.468018
ncvxqp3.mtx 2.6159286499 0.674072
para-4.mtx 2.81810760498 0.718994
nd3k.mtx 1.06000900269 0.48291
para-5.mtx 2.16507911682 0.629883
para-6.mtx 3.27706336975 0.88501
para-7.mtx 2.7961730957 0.830078
para-8.mtx 1.2309551239 0.401855
para-9.mtx 1.52277946472 0.448975
parabolic_fem.mtx 2.67601013184 0.712891
patents_main.mtx 0.918865203857 0.440918
pattern1.mtx 1.87110900879 0.716064
pcrystk03.mtx 2.63595581055 0.467041
pct20stif.mtx 4.18615341187 1.07104
pdb1HYS.mtx 0.972986221313 0.450928
pds-100.mtx 1.12915039062 0.445068
pds-30.mtx 39.8941040039 0.868164
pds-40.mtx 1.40309333801 0.510742
pds-50.mtx 1.96695327759 0.687012
pds-60.mtx 2.69913673401 0.47583

PageRank (with the same threshold)

Dataset Gunrock GSWITCH
rajat29.mtx 28.89585495 4.75415
TSOPF_FS_b162_c3.mtx 7.8558921814 3.32007
nw14.mtx 16.8740749359 4.60889
rajat21.mtx 28.9919376373 5.73608
pds-80.mtx 17.8439617157 3.71606
darcy003.mtx 8.00108909607 2.56592
mario002.mtx 8.11290740967 2.5
TSOPF_FS_b39_c19.mtx 11.5189552307 4.09473
TSOPF_RS_b39_c30.mtx 10.6821060181 4.12402
kneser_10_4_1.mtx 8.08501243591 3.38501
PR02R.mtx 9.53793525696 4.47803
c-big.mtx 50.0249862671 4.14014
net100.mtx 5.78093528748 2.83301
TSC_OPF_1047.mtx 7.62510299683 3.68579
pds-90.mtx 19.0241336823 4.09692
neos.mtx 85.8581066132 5.16602
nemsemm1.mtx 7.18688964844 3.52588
watson_1.mtx 14.6019458771 3.6189
wave.mtx 5.85389137268 2.85693
gupta1.mtx 12.9110813141 4.46509
pds-100.mtx 20.4699039459 4.58789
Freescale1.mtx 82.6079845428 8.70288
connectus.mtx 175.810098648 7.52905
helm2d03.mtx 8.12792778015 2.77075
TSOPF_FS_b162_c4.mtx 10.0297927856 4.271
pre2.mtx 10.7259750366 4.51782
ins2.mtx 23.0870246887 4.73511
cop20k_A.mtx 6.65092468262 3.18921
thermal2.mtx 28.3050537109 3.9978
soc-delicious.mtx 84.8710536957 6.69727
flickr.mtx 38.6641025543 8.73511
cage14.mtx 167.943954468 6.66504
circuit5M.mtx 202.360153198 13.2148
lp_osa_60.mtx 27.5950431824 6.17896
nlpkkt120.mtx 284.413814545 5.62891
inf-roadNet-PA.mtx 14.4050121307 4.22803
roadNet-PA.mtx 15.212059021 4.43896
ljournal-2008.mtx 464.427947998 20.8501
TSOPF_FS_b39_c30.mtx 16.5441036224 6.45288
inf-belgium_osm.mtx 16.9429779053 5.20801
belgium_osm.mtx 16.7000293732 5.20483
delaunay_n19.mtx 11.8069648743 4.46191
mc2depi.mtx 10.3089809418 3.44092
parabolic_fem.mtx 10.372877121 3.6189
Hamrle3.mtx 17.6539421082 3.771
Rucci1.mtx 309.336900711 8.98096
bibd_18_9.mtx 25.2511501312 7.11499
cont1_l.mtx 264.796972275 13.2761
karted.mtx 14.0228271484 6.66187
watson_2.mtx 27.8298854828 6.97583
roadNet-TX.mtx 19.1149711609 5.26807
GL7d22.mtx 41.2278175354 6.91699
GL7d16.mtx 52.521944046 6.99097
ecology2.mtx 13.9970779419 3.56006
ecology1.mtx 13.8919353485 3.53613
webbase-1M.mtx 100.652933121 10.6321
apache2.mtx 13.3981704712 3.62891
degme.mtx 61.3698959351 8.61133
debr.mtx 13.2520198822 4.20801
tmt_sym.mtx 13.5622024536 4.27905
largebasis.mtx 11.8281841278 4.87988
rt-retweet-crawl.mtx 48.8238334656 14.686
patents.mtx 119.863986969 15.157
GL7d19.mtx 228.003025055 8.22583
inf-netherlands_osm.mtx 24.1429805756 7.39893
netherlands_osm.mtx 24.1820812225 7.38794
in-2004.mtx 355.493068695 13.9128
inf-roadNet-CA.mtx 24.7769355774 7.44214
roadNet-CA.mtx 26.4029502869 7.46777
stat96v2.mtx 279.149055481 9.8269
nlpkkt80.mtx 64.6181106567 5.80811
soc-youtube-snap.mtx 200.951099396 15.3401
delaunay_n20.mtx 20.7948684692 8.32397
stat96v3.mtx 321.749210358 11.208
bibd_19_9.mtx 46.6759204865 13.7061
as-Skitter.mtx 114.364862442 18.2317
stormG2_1000.mtx 57.3320388794 13.7681
sls.mtx 426.111221313 26.646
relat9.mtx 705.684185028 39.0374
spal_004.mtx 55.732011795 11.0969
cit-Patents.mtx 164.915084839 27.137
rail4284.mtx 125.18119812 14.9351
12month1.mtx 44.4910526276 18.6318
web-wikipedia2009.mtx 241.136789322 31.002
soc-wiki-Talk-dir.mtx 452.569007874 32.8049
LargeRegFile.mtx 310.248851776 25.5569
tp-6.mtx 97.6030826569 23.9868
germany_osm.mtx 63.0660057068 19.457
delaunay_n21.mtx 39.6120548248 16.6882
hugetrace-00000.mtx 51.2290000916 20.564
inf-italy_osm.mtx 66.349029541 20.9329
venturiLevel3.mtx 52.4480342865 14.1841
inf-great-britain_osm.mtx 84.9270820618 26.0962
bibd_20_10.mtx 116.475105286 41.2791
hugetric-00000.mtx 64.2511844635 24.2322
bibd_22_8.mtx 108.223199844 46.418
inf-germany_osm.mtx 119.884967804 42.9219
delaunay_n22.mtx 76.9131183624 34.2512
inf-asia_osm.mtx 129.364013672 38.2571
adaptive.mtx 88.6628627777 28.127
delaunay_n23.mtx 150.447845459 71.8608

SSSP (Both enable stepping)

Dataset Gunrock GSWITCH
cfd2.mtx 45.0830459595 13.4231
net150.mtx 10.2601051331 3.89209
lhr71.mtx 23.5421657562 7.41211
lhr71c.mtx 23.6990451813 7.74316
pkustk03.mtx 26.0879993439 6.76514
inf-roadNet-PA.mtx 102.020980835 49.9529
rgg_n_2_18_s0.mtx 194.584136963 66.7258
inf-belgium_osm.mtx 213.896987915 68.147
belgium_osm.mtx 220.609191895 68.7439
sme3Dc.mtx 25.6111621857 10.96
mc2depi.mtx 292.74105835 136.824
parabolic_fem.mtx 279.83807373 134.624
3dtube.mtx 29.3908119202 7.77295
ohne2.mtx 17.8198814392 7.03003
Rucci1.mtx 20.4720497131 8.25586
m14b.mtx 27.8120040894 9.10498
sc-shipsec1.mtx 44.8811035156 14.2483
Dubcova3.mtx 55.9661407471 14.7678
web-arabic-2005.mtx 15.594959259 7.58398
oilpan.mtx 50.5058746338 15.481
cont1_l.mtx 18.6331272125 9.19482
n4c6-b8.mtx 9.05299186707 4.37183
C2000-9.mtx 5.29909133911 1.2251
GL7d14.mtx 9.80114936829 4.31787
s3dkt3m2.mtx 53.1549453735 24.0229
s4dkt3m2.mtx 57.5408935547 23.7241
bmwcra_1.mtx 35.8350296021 8.45801
barrier2-1.mtx 18.2540416718 8.97095
barrier2-3.mtx 18.7089443207 9.16895
n4c6-b9.mtx 8.53300094604 4.19897
barrier2-10.mtx 19.9751853943 9.36206
GL7d22.mtx 11.1780166626 5.31494
cant.mtx 94.7959442139 43.0439
GL7d16.mtx 14.4979953766 6.81494
offshore.mtx 24.3091583252 8.29321
pkustk10.mtx 28.3999443054 8.4292
t3dh.mtx 31.4099788666 8.22681
t3dh_a.mtx 31.3720703125 7.95605
t3dh_e.mtx 32.2780609131 8.24512
sc-shipsec5.mtx 60.8060379028 15.47
engine.mtx 29.7079086304 9.41504
s3dkq4m2.mtx 58.4588050842 26.5811
GL7d19.mtx 17.1990394592 7.89209
inf-netherlands_osm.mtx 271.272888184 88.019
netherlands_osm.mtx 275.174865723 86.9639
ch7-9-b5.mtx 11.3940238953 5.6377
pkustk11.mtx 44.8751449585 10.5911
sc-pkustk11.mtx 46.9770431519 10.636
ESOC.mtx 15.5799388885 6.63379
F2.mtx 52.1609802246 13.5662
af_4_k101.mtx 97.4349975586 32.6299
af_5_k101.mtx 100.863937378 31.613
af_2_k101.mtx 106.055023193 29.4373
af_3_k101.mtx 108.717918396 29.7341
af_0_k101.mtx 117.622138977 30.6531
af_1_k101.mtx 107.180831909 29.5391
stat96v2.mtx 16.6380405426 8.22705
nlpkkt80.mtx 36.0159873962 14.1489
af_shell1.mtx 114.135025024 42.6372
consph.mtx 37.5618934631 8.01904
gearbox.mtx 43.671131134 11.8269
pkustk13.mtx 54.4619560242 13.0132
sc-pkustk13.mtx 50.7719497681 12.5491
shipsec8.mtx 44.6410179138 14.0049
rgg_n_2_19_s0.mtx 297.246948242 118.626
boneS01.mtx 39.2539520264 11.948
auto.mtx 44.5890426636 17.7432
stat96v3.mtx 21.0981369019 8.51489
ch8-8-b5.mtx 14.51587677 6.88428
sls.mtx 40.8310890198 15.0789
IMDB.mtx 33.1969261169 11.989
misc-IMDB-bi.mtx 33.4169845581 11.8528
shipsec1.mtx 62.0079040527 16.804
ship_003.mtx 58.3879928589 13.1709
C4000-5.mtx 3.09491157532 1.39917
TF19.mtx 16.7171955109 7.86084
keller6.mtx 3.99589538574 1.57812
LargeRegFile.mtx 72.4780578613 23.988
MANN-a81.mtx 4.57000732422 1.69775
fcondp2.mtx 66.6508636475 23.0461
sc-pwtk.mtx 119.899032593 41.1472
fullb.mtx 60.9860420227 17.8699
troll.mtx 72.2188949585 18.3889
GL7d15.mtx 23.8001346588 8.15625
halfb.mtx 75.0980377197 24.3088
rgg_n_2_20_s0.mtx 606.960754395 239.585
inf-italy_osm.mtx 1217.22607422 452.819
pkustk14.mtx 61.0280036926 25.0701
frb100-40.mtx 5.99098205566 1.91406
inf-great-britain_osm.mtx 863.799072266 326.098
sc-msdoor.mtx 139.200210571 32.7151
inf-germany_osm.mtx 720.188110352 345.122
inf-asia_osm.mtx 4967.0078125 1966.74
rgg_n_2_21_s0.mtx 1799.15112305 551.594
sc-ldoor.mtx 269.086120605 87.2778
soc-pokec.mtx 96.5909957886 29.4673
socfb-A-anon.mtx 68.3159790039 30.3987
soc-livejournal.mtx 164.842132568 64.8552
rgg_n_2_22_s0.mtx 4406.23193359 1156.96
channel-500x100x100-b050.mtx 992.661010742 473.475
ca-hollywood-2009.mtx 80.6810836792 30.4668

BC

Dataset Gunrock GSWITCH
dbir1.mtx 5.64908981323 1.6298832
dbic1.mtx 3.51595878601 1.66967752
pkustk05.mtx 17.8558826447 8.82495104
web-NotreDame.mtx 10.4839801788 3.9970712
pwtk.mtx 39.4740104675 11.32226602
poisson3Db.mtx 10.9560489655 3.97412198
Ge99H100.mtx 6.79111480713 2.37621998
ASIC_320k.mtx 4.27484512329 1.2172857
EternityII_Etilde.mtx 3.59201431274 1.13916068
TSOPF_RS_b2383.mtx 27.7318954468 5.89501804
socfb-UGA50.mtx 3.44514846802 0.86987294
TSOPF_RS_b2383_c1.mtx 24.5890617371 5.52734434
Maragal_7.mtx 4.39596176147 0.96167048
pkustk07.mtx 6.69503211975 2.16308654
tsyl201.mtx 16.8790817261 6.10522434
net125.mtx 4.46820259094 1.63793914
std1_Jac2.mtx 5.44810295105 1.7211905
socfb-UIllinois.mtx 3.77702713013 0.9782716
socfb-UIllinois20.mtx 3.47089767456 1.01000974
Ga19As19H42.mtx 7.13896751404 2.61865254
filter3D.mtx 21.9359397888 7.38500868
n4c6-b7.mtx 5.27501106262 1.282959
pct20stif.mtx 11.7959976196 5.42016614
relat8.mtx 5.5980682373 1.54345668
ramage02.mtx 8.23211669922 3.17724462
TSOPF_RS_b300_c1.mtx 4.15110588074 1.1938481
socfb-UF.mtx 4.5120716095 0.90991174
Si41Ge41H72.mtx 6.89101219177 2.39965928
cfd2.mtx 30.1990509033 8.02783178
net150.mtx 4.63700294495 1.50097574
lp_nug30.mtx 3.8959980011 1.21753054
ch7-9-b4.mtx 4.95600700378 1.45288092
socfb-Texas84.mtx 3.58700752258 0.95776304
ohne2.mtx 12.9070281982 5.61621094
pkustk08.mtx 7.99179077148 2.87622114
nd3k.mtx 4.87613677979 1.7653815
m14b.mtx 13.5400295258 5.77124004
av41092.mtx 6.42108917236 1.321044
sc-shipsec1.mtx 24.2938995361 7.90893488
Dubcova3.mtx 21.6100215912 8.4899902
web-arabic-2005.mtx 15.095949173 6.94921724
oilpan.mtx 27.3699760437 9.9147953
cont1_l.mtx 7.31897354126 1.6982423
karted.mtx 3.61394882202 0.79248
IG5-18.mtx 3.60298156738 1.04394538
n4c6-b8.mtx 5.25689125061 1.18066448
bmwcra_1.mtx 15.4309272766 7.53588798
barrier2-1.mtx 9.06801223755 3.98217694
barrier2-2.mtx 8.86702537537 3.8205567
barrier2-3.mtx 11.1320018768 3.86108354
watson_2.mtx 6.77895545959 2.86108384
ch8-8-b4.mtx 5.14698028564 1.55029302
barrier2-10.mtx 12.4750137329 4.0341804
barrier2-12.mtx 9.49597358704 3.98779214
Chebyshev4.mtx 2.86889076233 0.95263668
offshore.mtx 17.2069072723 6.1435549
soc-BlogCatalog.mtx 3.97491455078 1.12866188
debr.mtx 10.0839138031 3.9633791
pkustk10.mtx 18.6841487885 9.11206088
t3dh_e.mtx 15.6941413879 6.65014548
soc-LiveMocha.mtx 4.74786758423 1.04516638
sc-shipsec5.mtx 30.935049057 9.37084984
rt-retweet-crawl.mtx 9.81593132019 3.5507808
s3dkq4m2.mtx 51.2230377197 10.87500028
human_gene1.mtx 5.10787963867 1.6193855
amazon0505.mtx 11.3439559937 3.7346184
amazon0601.mtx 8.99696350098 4.18579184
kron_g500-logn16.mtx 4.16588783264 0.86743114
Si34H36.mtx 10.3950500488 3.16625954
SiO2.mtx 5.22112846375 2.07275394
ESOC.mtx 8.79788398743 1.76245124
F2.mtx 14.6589279175 6.55444344
GL7d23.mtx 6.34598731995 1.6284192
stat96v2.mtx 6.29806518555 2.07128928
Trec14.mtx 4.20904159546 1.1772462
consph.mtx 15.319108963 7.22705114
soc-youtube-snap.mtx 7.94196128845 3.80932582
Ga10As10H30.mtx 8.51392745972 3.65698198
soc-flickr.mtx 8.76212120056 3.0458979
soc-FourSquare.mtx 5.12504577637 1.21557624
auto.mtx 19.0489292145 9.4819331
stat96v3.mtx 5.7520866394 2.2043459
spal_004.mtx 4.60600852966 1.88305734
IMDB.mtx 10.3280544281 3.01562376
misc-IMDB-bi.mtx 9.17220115662 2.99853554
shipsec1.mtx 25.7458686829 9.42944314
12month1.mtx 8.73398780823 3.120849
human_gene2.mtx 3.87001037598 1.10302724
web-Google.mtx 10.5278491974 5.1472154
soc-lastfm.mtx 8.79693031311 3.0751958
gupta3.mtx 4.83584403992 1.06494174
mip1.mtx 3.43608856201 0.79955956
tp-6.mtx 8.48698616028 2.9890133
fcondp2.mtx 48.4108924866 10.4499519
fullb.mtx 29.205083847 12.93554564
troll.mtx 30.5089950562 11.73388634
soc-digg.mtx 8.52608680725 3.0615233
GL7d15.mtx 8.77714157104 2.07202154
halfb.mtx 44.6090698242 11.2448724
pkustk14.mtx 39.2370223999 15.11206
soc-flixster.mtx 8.38685035706 3.0053704

V100

DOBFS (both without tuning parameters)

Dataset Gunrock GSWITCH
FullChip.mtx 0.810146331787 0.220947
Freescale1.mtx 1.44505500793 0.181885
BenElechi1.mtx 33.2617759705 5.65405
Ill_Stokes.mtx 11.9299888611 1.47095
ABACUS_shell_hd.mtx 37.3270492554 5.83008
ABACUS_shell_ld.mtx 37.7168655396 5.87207
ABACUS_shell_md.mtx 37.4908447266 5.83521
ABACUS_shell_ud.mtx 37.887096405 6.01904
Kemelmacher.mtx 2.35390663147 0.338135
OPF_6000.mtx 19.143819809 1.51196
Na5.mtx 1.78503990173 0.499023
bcsstm36.mtx 20.1640129089 1.6521
Kuu.mtx 12.0270252228 1.42285
bcsstk38.mtx 20.9591388702 2.12695
FEM_3D_thermal1.mtx 26.1988639832 2.98218
2D_27628_bjtcai.mtx 30.620098114 3.97095
Ga41As41H72.mtx 6.74986839294 0.684082
OPF_10000.mtx 18.7091827393 2.39307
ASIC_680k.mtx 4.14800643921 0.317871
c-66b.mtx 8.2471370697 1.47412
GT01R.mtx 18.8570022583 4.96387
c-62ghs.mtx 1.16896629333 0.375
FEM_3D_thermal2.mtx 69.6089248657 9.01221
G2_circuit.mtx 32.3638916016 4.39185
3D_28984_Tetra.mtx 2.81691551208 0.856201
EAT_SR.mtx 1.34992599487 0.468018
EAT_RS.mtx 1.30009651184 0.446045
HTC_336_9129.mtx 11.2030506134 1.69482
Baumann.mtx 46.2191085815 5.63599
HTC_336_4438.mtx 12.9380226135 1.85815
Franz11.mtx 1.26004219055 0.345947
Andrews.mtx 3.78394126892 0.50708
Pres_Poisson.mtx 28.4960269928 2.64307
HEP-th-new.mtx 2.11906433105 0.724121
GL7d13.mtx 1.62696838379 0.472168
Si5H12.mtx 10.5011463165 1.0249
ia-wiki-Talk.mtx 1.4820098877 0.607178
RFdevice.mtx 2.49290466309 0.404053
delaunay_n17.mtx 36.6439819336 4.67896
soc-Epinions1.mtx 1.73401832581 0.631836
fe-ocean.mtx 40.3671264648 5.2439
ASIC_100k.mtx 1.31797790527 0.375244
Si10H16.mtx 4.19187545776 0.584961
fe-tooth.mtx 12.540102005 1.66211
2D_54019_highK.mtx 25.2871513367 2.43311
Dubcova2.mtx 19.7470188141 2.5022
3Dspectralwave2.mtx 6.92510604858 0.85791
3D_51448_3D.mtx 67.4510040283 2.09204
hamming10-2.mtx 0.686168670654 0.185059
Raj1.mtx 54.7978897095 3.23901
MANN-a45.mtx 0.661134719849 0.143066
Maragal_6.mtx 0.903129577637 0.211914
p-hat1500-2.mtx 0.653982162476 0.162842
frb50-23-3.mtx 0.703096389771 0.168945
frb50-23-2.mtx 0.695943832397 0.165039
frb50-23-1.mtx 0.686883926392 0.164062
frb50-23-5.mtx 0.730037689209 0.166748
IG5-16.mtx 1.92809104919 0.491211
GL7d24.mtx 1.92403793335 0.545898
tech-caidaRouterLevel.mtx 6.73198699951 1.08521
SiO.mtx 3.37195396423 0.670898
fe_rotor.mtx 14.6219730377 1.96094
frb53-24-4.mtx 0.771999359131 0.166748
frb53-24-2.mtx 0.702857971191 0.169189
frb53-24-1.mtx 0.680923461914 0.166992
frb53-24-5.mtx 0.726938247681 0.168213
frb53-24-3.mtx 0.718116760254 0.166992
ca-dblp-2010.mtx 5.4759979248 0.812012
rgg_n_2_17_s0.mtx 75.9391784668 12.582
598a.mtx 10.7741355896 1.48511
NotreDame_www.mtx 33.7839126587 1.94214
Lin.mtx 47.8489379883 5.59277
2cubes_sphere.mtx 10.8969211578 1.57593
EternityII_A.mtx 1.36303901672 0.352051
CO.mtx 3.45611572266 0.886963
delaunay_n18.mtx 48.4330673218 5.79004
coAuthorsCiteseer.mtx 4.42790985107 1.17285
ca-citeseer.mtx 4.50396537781 1.10205
ca-MathSciNet.mtx 4.15802001953 0.926025
ASIC_680ks.mtx 10.9958648682 1.40283
p-hat1500-3.mtx 0.709056854248 0.169922
Ge87H76.mtx 7.00497627258 0.874023
HFE18_96_in.mtx 5.08284568787 0.478027
PR02R.mtx 80.5041809082 11.4221
C2000-5.mtx 0.684022903442 0.1521
af_shell2.mtx 69.4530029297 8.31201
atmosmodd.mtx 95.1538085938 6.81592
atmosmodj.mtx 95.1681137085 6.63086
bcsstk39.mtx 123.610977173 10.3391
IG5-17.mtx 2.12693214417 0.461914
frb59-26-4.mtx 0.772953033447 0.156006
frb59-26-1.mtx 0.734806060791 0.182861
frb59-26-2.mtx 0.702142715454 0.169189
frb59-26-3.mtx 0.802040100098 0.176025
frb59-26-5.mtx 0.715017318726 0.157715
ca-dblp-2012.mtx 5.8650970459 0.903076
144.mtx 10.1239681244 1.30615
H2O.mtx 8.92281532288 1.22876
Hook_1498.mtx 17.893075943 2.61499
connectus.mtx 1.87683105469 0.384033
Ge99H100.mtx 4.99701499939 1.06909

CC (Both are edge-centric)

Dataset Gunrock GSWITCH
ASIC_320k.mtx 2.0158290863 0.724854
adaptive.mtx 0.838041305542 0.397949
delaunay_n22.mtx 1.49202346802 0.323975
598a.mtx 0.791072845459 0.321045
delaunay_n23.mtx 0.75101852417 0.331055
auto.mtx 0.810146331787 0.325684
delaunay_n24.mtx 0.80394744873 0.322021
bio-pdb1HYS.mtx 0.591993331909 0.215088
ASIC_680k.mtx 2.09784507751 0.75708
c-62ghs.mtx 1.51610374451 0.555908
fe-ocean.mtx 0.939130783081 0.318848
c-66b.mtx 0.911951065063 0.397217
fe-tooth.mtx 0.854969024658 0.328125
channel-500x100x100-b050.mtx 0.894069671631 0.373047
citationCiteseer.mtx 0.746965408325 0.209961
Andrews.mtx 0.881910324097 0.273193
cnr-2000.mtx 1.2149810791 0.388916
fe_rotor.mtx 2.07591056824 0.779297
co-papers-citeseer.mtx 1.99413299561 0.351074
co-papers-dblp.mtx 1.09601020813 0.412842
Baumann.mtx 0.710964202881 0.217773
coAuthorsCiteseer.mtx 1.59192085266 0.549072
Dubcova3.mtx 1.21808052063 0.305908
connectus.mtx 1.53207778931 0.429199
consph.mtx 1.26004219055 0.576904
hugetrace-00000.mtx 0.960826873779 0.213867
m14b.mtx 1.22714042664 0.320068
cop20k_A.mtx 1.2481212616 0.318115
hugetrace-00010.mtx 0.837087631226 0.340088
delaunay_n17.mtx 2.75492668152 0.549805
Dubcova2.mtx 0.624179840088 0.188965
hugetrace-00020.mtx 2.98810005188 0.573975
delaunay_n18.mtx 1.45602226257 0.319092
delaunay_n19.mtx 1.16491317749 0.364746
ca-MathSciNet.mtx 1.22094154358 0.496826
hugetric-00000.mtx 1.02210044861 0.201904
ca-citeseer.mtx 1.0290145874 0.306885
delaunay_n20.mtx 1.1899471283 0.311279
ca-coauthors-dblp.mtx 1.18494033813 0.246826
C2000-5.mtx 1.08194351196 0.35791
hugetric-00010.mtx 1.42908096313 0.662109
C2000-9.mtx 2.15816497803 0.703125
hugetric-00020.mtx 1.14703178406 0.546875
C4000-5.mtx 0.857830047607 0.224121
inf-asia_osm.mtx 1.32489204407 0.490967
MANN-a45.mtx 1.54185295105 0.657227
inf-belgium_osm.mtx 2.03990936279 0.343994
MANN-a81.mtx 1.00994110107 0.322754
inf-europe_osm.mtx 1.41000747681 0.430176
hamming10-2.mtx 1.28602981567 0.396973
inf-germany_osm.mtx 1.66487693787 0.39502
keller6.mtx 0.946998596191 0.237061
EAT_RS.mtx 0.843048095703 0.156982
144.mtx 1.6930103302 0.688965
delaunay_n21.mtx 0.617980957031 0.185791
inf-great-britain_osm.mtx 0.849008560181 0.197021
inf-italy_osm.mtx 0.639915466309 0.166748
inf-netherlands_osm.mtx 0.810861587524 0.196045
inf-road_central.mtx 0.875949859619 0.206787
inf-road_usa.mtx 0.931978225708 0.205078
kron_g500-logn16.mtx 0.88095664978 0.189209
kron_g500-logn17.mtx 0.913858413696 0.202881
kron_g500-logn18.mtx 0.677108764648 0.224121
kron_g500-logn19.mtx 0.790119171143 0.225098
kron_g500-logn20.mtx 1.25503540039 0.37915
kron_g500-logn21.mtx 1.08599662781 0.339844
EAT_SR.mtx 1.0302066803 0.328125
Lin.mtx 1.65104866028 0.521973
packing-500x100x100-b050.mtx 1.13701820374 0.327881
FEM_3D_thermal1.mtx 1.69610977173 0.436035
rgg_n_2_17_s0.mtx 0.564098358154 0.181885
FEM_3D_thermal2.mtx 0.916004180908 0.208008
rgg_n_2_18_s0.mtx 0.919103622437 0.199951
Fault_639.mtx 0.929117202759 0.203125
RM07R.mtx 4.28700447083 0.816162
bcsstk38.mtx 3.50999832153 0.673096
Raj1.mtx 2.19392776489 0.694092
av41092.mtx 1.29294395447 0.354248
Rucci1.mtx 1.20806694031 0.431152
barrier2-1.mtx 1.35684013367 0.380859
Serena.mtx 3.27110290527 0.799072
barrier2-10.mtx 2.88200378418 0.521973
Si10H16.mtx 1.02496147156 0.38501
barrier2-11.mtx 2.18176841736 0.497803
Si34H36.mtx 7.32278823853 1.39282
barrier2-12.mtx 8.65793228149 2.23901
barrier2-2.mtx 11.0530853271 4.13599
Si5H12.mtx 2.47001647949 0.529053
barrier2-3.mtx 8.74900817871 2.43896
Si87H76.mtx 12.3949050903 4.39795
bcsstk39.mtx 10.6699466705 2.69189
SiO.mtx 15.2499675751 5.20605
bcsstm36.mtx 4.57191467285 0.822998
SiO2.mtx 17.3659324646 5.63599
barrier2-4.mtx 18.3839797974 6.68921
Stanford.mtx 11.885881424 2.36694
af_shell3.mtx 19.8881626129 7.58472
Stanford_Berkeley.mtx 17.0118808746 3.08911
barrier2-9.mtx 7.67803192139 2.00488
StocF-1465.mtx 5.51986694336 1.06592
bibd_17_8.mtx 5.52201271057 1.06592

PageRank (with the same threshold)

Dataset Gunrock GSWITCH
aft02.mtx 8.52394104004 1.15112
bauru5727.mtx 15.0609016418 1.177
BenElechi1.mtx 5.76496124268 1.06592
Ill_Stokes.mtx 6.70409202576 1.16309
ABACUS_shell_hd.mtx 6.28113746643 1.06982
ABACUS_shell_ld.mtx 6.29806518555 1.04492
ABACUS_shell_md.mtx 6.26993179321 1.0769
ABACUS_shell_ud.mtx 6.27899169922 1.07397
Kemelmacher.mtx 8.07285308838 1.271
airfoil_2d.mtx 11.6350650787 1.10107
OPF_6000.mtx 7.69400596619 1.21094
Stanford_Berkeley.mtx 245.684862137 1.46094
Na5.mtx 7.50398635864 1.28711
bcsstm36.mtx 6.50191307068 1.10522
Kuu.mtx 6.58011436462 1.09717
bcsstk38.mtx 6.94894790649 1.18896
TSOPF_RS_b162_c1.mtx 11.4090442657 1.43408
FEM_3D_thermal1.mtx 7.11894035339 1.15088
2D_27628_bjtcai.mtx 7.45511054993 1.23486
Ga41As41H72.mtx 7.70688056946 1.22998
OPF_10000.mtx 8.19993019104 1.32202
ASIC_680k.mtx 20.4410552979 4.22314
c-66b.mtx 4.72807884216 1.62695
GT01R.mtx 6.78396224976 1.11694
ASIC_100ks.mtx 18.8179016113 1.97778
c-62ghs.mtx 3.9439201355 1.72021
FEM_3D_thermal2.mtx 26.5409946442 1.2251
G2_circuit.mtx 11.0960006714 1.15894
3D_28984_Tetra.mtx 8.71610641479 1.47412
TSOPF_FS_b162_c1.mtx 13.8349533081 1.51782
EAT_SR.mtx 10.575056076 1.29712
EAT_RS.mtx 4.0340423584 1.31421
Baumann.mtx 10.7297897339 1.08813
HTC_336_9129.mtx 38.8889312744 2.13501
Franz11.mtx 9.99689102173 1.49707
HTC_336_4438.mtx 43.6861515045 2.23389
Andrews.mtx 3.21316719055 1.5293
Pres_Poisson.mtx 7.82108306885 1.23828
HEP-th-new.mtx 11.8310451508 1.41992
GL7d13.mtx 11.1351013184 1.40503
Si5H12.mtx 11.2309455872 1.39673
ia-wiki-Talk.mtx 13.839006424 1.5979
TSOPF_FS_b39_c7.mtx 14.848947525 2.16235
RFdevice.mtx 10.1978778839 1.823
TSOPF_FS_b300_c3.mtx 17.3480510712 1.48511
delaunay_n17.mtx 13.1649971008 1.3501
TSC_OPF_300.mtx 13.1871700287 1.57593
soc-Epinions1.mtx 15.5568122864 2.20972
fe-ocean.mtx 3.36503982544 1.22998
FullChip.mtx 247.121095657 2.52905
Si10H16.mtx 11.4130973816 1.6499
fe-tooth.mtx 6.95300102234 1.59204
2D_54019_highK.mtx 9.14692878723 1.45898
Dubcova2.mtx 3.98898124695 1.54907
GL7d18.mtx 147.60684967 2.05786
3Dspectralwave2.mtx 10.38813591 1.29688
3D_51448_3D.mtx 9.6230506897 1.68896
F1.mtx 9.80281829834 1.49414
hamming10-2.mtx 4.85610961914 1.60498
Raj1.mtx 41.944026947 3.13599
MANN-a45.mtx 3.91697883606 1.5293
Maragal_6.mtx 12.2179985046 2.16797
GL7d20.mtx 177.146911621 2.23486
frb50-23-3.mtx 3.65900993347 1.61084
TF17.mtx 24.7800350189 1.89893
TSOPF_RS_b162_c3.mtx 16.1099433899 2.17969
bas1lp.mtx 19.5019245148 1.78223
IG5-16.mtx 13.9648914337 2.21387
GL7d24.mtx 15.9640312195 1.84009
tech-caidaRouterLevel.mtx 22.4528312683 2.97192
SiO.mtx 15.408039093 2.09497
bibd_17_8.mtx 27.8360843658 5.51709
frb53-24-1.mtx 3.30281257629 1.59521
TSOPF_RS_b2052_c1.mtx 19.7908878326 1.55396
rgg_n_2_17_s0.mtx 18.1479454041 2.31519
598a.mtx 5.50603866577 1.97485
ASIC_320ks.mtx 46.4618206024 3.80005
Lin.mtx 21.7969417572 1.51294
NotreDame_www.mtx 142.706155777 2.51807
2cubes_sphere.mtx 15.7780647278 1.81201
TSOPF_RS_b162_c4.mtx 18.3827877045 2.59497
EternityII_A.mtx 19.9248790741 4.18506
CO.mtx 4.95314598083 1.90015
delaunay_n18.mtx 11.0061168671 1.81104
ca-citeseer.mtx 22.8860378265 2.6958
ca-MathSciNet.mtx 32.9921245575 2.41895
ASIC_680ks.mtx 54.0509223938 3.34521
p-hat1500-3.mtx 6.38294219971 1.8501
TSOPF_FS_b162_c3.mtx 20.6921100616 2.7749
HFE18_96_in.mtx 11.048078537 1.875
TSOPF_FS_b39_c19.mtx 24.6889591217 4.08472
PR02R.mtx 43.2641506195 1.75684
C2000-5.mtx 5.07307052612 2.05029
TSC_OPF_1047.mtx 20.6589698792 2.47095
atmosmodd.mtx 29.4840335846 1.63696
atmosmodj.mtx 29.0579795837 1.64014
bcsstk39.mtx 11.7499828339 1.81787
af_shell2.mtx 11.7619037628 1.70483
IG5-17.mtx 21.1918354034 2.91187
frb59-26-1.mtx 4.04095649719 1.91602
ca-dblp-2012.mtx 30.7829380035 4.05005

SSSP (Both enable stepping)

Dataset Gunrock GSWITCH
cit-Patents.mtx 0.473022460938 0.159912
flickr.mtx 0.620126724243 0.146973
FullChip.mtx 0.59986114502 0.278076
hood.mtx 0.715017318726 0.145996
Freescale1.mtx 1.01399421692 0.362061
ex35.mtx 19.6208953857 7.1189
gyro_m.mtx 23.6229896545 5.25806
aft02.mtx 19.6800231934 6.12988
bauru5727.mtx 6.67500495911 2.1748
cavity25.mtx 13.986825943 2.9043
chem_master1.mtx 40.2240753174 15.4421
ex19.mtx 20.3280448914 6.60205
Ill_Stokes.mtx 7.47799873352 2.39624
ABACUS_shell_hd.mtx 26.1940956116 11.2441
ABACUS_shell_ld.mtx 26.8030166626 12.0798
ABACUS_shell_md.mtx 27.7600288391 11.3792
ABACUS_shell_ud.mtx 26.2751579285 11.574
epb2.mtx 16.1299705505 6.37109
Kemelmacher.mtx 3.16095352173 0.827148
bcsstk28.mtx 11.519908905 3.13013
hvdc2.mtx 15.212059021 3.32715
bcsstk25.mtx 15.1059627533 7.02905
airfoil_2d.mtx 10.4382038116 4.43896
c-60.mtx 3.32403182983 1.49487
chipcool1.mtx 9.04583930969 3.03809
c-61.mtx 3.18694114685 0.994141
OPF_6000.mtx 7.45797157288 2.3501
Na5.mtx 4.97388839722 2.33813
bcircuit.mtx 17.3900127411 6.75195
bcsstm36.mtx 8.9750289917 2.87793
Kuu.mtx 10.4730129242 2.95386
c-53.mtx 4.88495826721 1.80103
c-56.mtx 5.39088249207 1.97974
bcsstk38.mtx 13.2610797882 5.84009
ecl32.mtx 12.1018886566 3.65625
c-54.mtx 5.11908531189 2.19995
c-57.mtx 3.6768913269 1.61914
c-55.mtx 5.15604019165 2.3811
FEM_3D_thermal1.mtx 11.8019580841 5.72803
2D_27628_bjtcai.mtx 19.2420482635 8.77783
dawson5.mtx 21.8060016632 4.99976
Ga41As41H72.mtx 4.98294830322 1.52905
OPF_10000.mtx 12.0000839233 4.30298
goodwin.mtx 22.3550796509 5.01904
ASIC_680k.mtx 3.00312042236 1.40601
c-66b.mtx 4.63199615479 2.25781
c-66.mtx 5.08499145508 2.24683
ex40.mtx 13.4019851685 4.02905
af23560.mtx 21.8648910522 9.58594
GT01R.mtx 17.914056778 7.46899
deltaX.mtx 3.87096405029 1.18628
epb3.mtx 57.0840835571 21.5391
c-58.mtx 5.4759979248 2.11597
c-62ghs.mtx 4.54902648926 1.57495
c-62.mtx 4.47106361389 1.73999
cage11.mtx 7.44104385376 2.125
FEM_3D_thermal2.mtx 39.019821167 14.6091
bayer01.mtx 6.57916069031 1.85596
G2_circuit.mtx 23.7340927124 7.62402
c-70.mtx 4.47702407837 2.04004
TSOPF_FS_b162_c1.mtx 3.6940574646 1.74902
c-72.mtx 5.863904953 2.37183
Baumann.mtx 29.9851894379 9.16406
HTC_336_4438.mtx 7.2660446167 3.4397
Franz11.mtx 4.46701049805 1.43701
Andrews.mtx 5.8159828186 1.79419
Pres_Poisson.mtx 12.6309394836 5.69507
HEP-th-new.mtx 5.18703460693 2.05396
GL7d13.mtx 5.02181053162 1.62109
Si5H12.mtx 6.88576698303 2.9668
ia-wiki-Talk.mtx 4.14896011353 1.88306
TSOPF_FS_b39_c7.mtx 3.34811210632 1.66016
RFdevice.mtx 4.81200218201 1.54907
dc3.mtx 2.94089317322 1.30786
dc2.mtx 3.00693511963 1.30371
dc1.mtx 5.14101982117 1.48804
bundle1.mtx 4.0431022644 1.45605
c-71.mtx 6.43110275269 3.02368
delaunay_n17.mtx 27.2860527039 10.0618
fe-ocean.mtx 27.2579193115 9.72388
cit-HepPh.mtx 10.272026062 2.23315
e40r0100.mtx 17.4868106842 6.80908
2D_54019_highK.mtx 15.0229930878 6.23291
Dubcova2.mtx 18.1288719177 6.85596
GL7d18.mtx 4.74810600281 1.88208
3Dspectralwave2.mtx 10.3261470795 2.57788
gyro.mtx 38.7840270996 9.81592
gyro_k.mtx 37.2538566589 11.5979
ibm_matrix_2.mtx 19.6359157562 5.06592
case39.mtx 5.54203987122 1.73901
fp.mtx 7.15494155884 2.43677
MANN-a45.mtx 2.01797485352 0.700928
denormal.mtx 43.0009384155 18.7329
ex11.mtx 12.8948688507 5.95288
GL7d20.mtx 5.03897666931 2.01587
c-73.mtx 5.70917129517 2.21704
c-73b.mtx 5.99884986877 2.21704
bcsstk37.mtx 15.457868576 6.02295
boyd1.mtx 3.12495231628 1.10693
bcsstk36.mtx 10.9529495239 5.13916
bcsstk31.mtx 17.6320075989 7.39917

BC

Dataset Gunrock GSWITCH
Freescale1.mtx 1.2309551239 0.602295
bauru5727.mtx 9.52816009521 5.09521
Ill_Stokes.mtx 8.72302055359 4.63086
ABACUS_shell_hd.mtx 33.0529212952 18.6753
ABACUS_shell_ld.mtx 33.0369491577 18.9153
ABACUS_shell_md.mtx 33.3678741455 18.79
ABACUS_shell_ud.mtx 33.5099716187 18.7532
Kemelmacher.mtx 2.23803520203 1.00415
OPF_6000.mtx 8.86607170105 4.84473
Na5.mtx 2.52294540405 1.28003
Kuu.mtx 9.05394554138 4.28784
FEM_3D_thermal1.mtx 16.608953476 10.811
2D_27628_bjtcai.mtx 21.5289592743 12.8762
Ga41As41H72.mtx 3.55887413025 2.06934
OPF_10000.mtx 13.5018825531 7.77783
c-66b.mtx 3.02195549011 1.61084
c-62ghs.mtx 2.36797332764 0.901123
bayer01.mtx 4.44078445435 2.17651
G2_circuit.mtx 26.535987854 13.95
3D_28984_Tetra.mtx 3.23390960693 1.77319
EAT_SR.mtx 2.97689437866 0.99707
EAT_RS.mtx 3.014087677 0.995117
Baumann.mtx 36.7031097412 20.2581
Franz11.mtx 2.77805328369 0.94751
Andrews.mtx 4.28485870361 1.49927
HEP-th-new.mtx 4.70209121704 1.59937
Si5H12.mtx 4.0910243988 2.47827
ia-wiki-Talk.mtx 3.73101234436 1.44507
TSOPF_FS_b39_c7.mtx 3.06296348572 1.71313
RFdevice.mtx 4.0500164032 1.22095
delaunay_n17.mtx 28.5160541534 15.46
TSC_OPF_300.mtx 4.87995147705 2.44482
soc-Epinions1.mtx 3.33499908447 1.66089
fe-ocean.mtx 29.6120643616 17.0408
Si10H16.mtx 3.67498397827 1.73608
fe-tooth.mtx 9.70697402954 5.29004
2D_54019_highK.mtx 12.7558708191 7.66235
Dubcova2.mtx 13.2129192352 8.11523
GL7d18.mtx 3.96394729614 2.17114
3Dspectralwave2.mtx 3.99780273438 2.55615
3D_51448_3D.mtx 4.90689277649 2.67407
hamming10-2.mtx 1.56903266907 0.484863
fp.mtx 10.6751918793 5.79907
GL7d20.mtx 4.03308868408 2.12012
boyd1.mtx 5.55300712585 3.64795
p-hat1500-2.mtx 2.01416015625 0.64502
frb50-23-3.mtx 1.9519329071 0.628662
frb50-23-2.mtx 1.60479545593 0.615723
frb50-23-4.mtx 1.65176391602 0.62085
frb50-23-1.mtx 1.65414810181 0.589355
frb50-23-5.mtx 1.61004066467 0.645264
TF17.mtx 5.39612770081 1.9126
TSOPF_RS_b162_c3.mtx 3.07393074036 1.4043
bas1lp.mtx 2.35915184021 1.36108
IG5-16.mtx 3.0369758606 1.17285
tech-caidaRouterLevel.mtx 5.68699836731 2.79297
TSOPF_RS_b39_c19.mtx 3.04102897644 1.19995
SiO.mtx 3.48210334778 1.9668
fe_rotor.mtx 10.9360218048 6.21118
frb53-24-4.mtx 1.68800354004 0.5979
frb53-24-2.mtx 1.64890289307 0.586182
frb53-24-5.mtx 2.07281112671 0.572998
ca-dblp-2010.mtx 5.774974823 2.61694
598a.mtx 8.24880599976 4.63599
NotreDame_www.mtx 7.60817527771 4.60913
Lin.mtx 35.6369018555 17.6152
2cubes_sphere.mtx 9.03701782227 4.26904
TSOPF_RS_b162_c4.mtx 3.30805778503 1.26685
EternityII_A.mtx 4.0819644928 1.63013
CO.mtx 6.12211227417 2.5481
delaunay_n18.mtx 35.2849960327 17.811
coAuthorsCiteseer.mtx 7.88593292236 3.13306
ca-citeseer.mtx 7.62414932251 3.11304
ca-MathSciNet.mtx 5.64098358154 2.80811
ASIC_680ks.mtx 8.12411308289 3.9707
p-hat1500-3.mtx 1.67894363403 0.654297
Ge87H76.mtx 5.38897514343 2.42603
amazon0302.mtx 7.94291496277 3.61108
HFE18_96_in.mtx 3.20982933044 1.14478
TSOPF_FS_b39_c19.mtx 3.56888771057 1.96289
TSC_OPF_1047.mtx 4.71711158752 2.73389
atmosmodd.mtx 40.1320457458 20.7468
atmosmodj.mtx 39.7701263428 20.9058
frb59-26-1.mtx 2.10285186768 0.669678
frb59-26-2.mtx 2.09093093872 0.670898
frb59-26-3.mtx 2.13599205017 0.61499
frb59-26-5.mtx 2.09212303162 0.681152
ca-dblp-2012.mtx 6.46686553955 2.68506
144.mtx 9.45997238159 3.93896
H2O.mtx 7.9460144043 3.67798
connectus.mtx 4.95100021362 1.42334
Ge99H100.mtx 6.46305084229 2.98511
citationCiteseer.mtx 8.00800323486 3.26562
ASIC_320k.mtx 4.5211315155 1.93164
EternityII_Etilde.mtx 3.19790840149 1.15698
ins2.mtx 7.33709335327 2.77271
Maragal_8.mtx 3.39698791504 1.38501
Si41Ge41H72.mtx 6.56604766846 2.60571
NotreDame_actors.mtx 5.5079460144 2.5271
inf-roadNet-PA.mtx 90.6808395386 43.26
rgg_n_2_18_s0.mtx 79.6859283447 50.6082

License

All the libraryies, examples, and source codes of GSWITCH are released under Apache 2.0.