add baseline

README.md

@@ -12,18 +12,38 @@ This artifact supports evaluation claims in the paper #18 of OOPSLA'24 round 1:
   1. The format conversion and compute kernels generated by UniSparse achieve performance that matches prior MLIR SparseTensor and TACO compilers, while provides a broader coverage for handling a wider range of custom formats. (Section 7.3 and 7.4)
   2. The adoption of custom formats enabled by UniSparse leads to improved performance for common sparse linear algebra kernels on hardware platforms such as an Intel multi-core CPU, an NVIDIA GPU and a simulated PIM device. (Section 7.2)
 
-In the Step-by-Step Instructions section below, we provide instructions on how to generate performance numbers in Section 7.2, 7.3 and 7.4 of the paper that support the claims above. 
+In the Step-by-Step Instructions section below, we provide instructions for 4 experiments. Experiment 1 and 2 reproduces the performance numbers in Section 7.3 and 7.4 respectively, which support the claim 1. Experiment 3 reproduces the performance numbers in Section 7.2.1, which supports claim 2. Experiment 4 reproduces the performance numbers in Section 7.2.2, which also supports claim 2. Experiment 4 is optional depending on whether the evaluator has an NVIDIA GPU and CUDA version >= 11.0. We use 4 datasets for each experiment, as more datasets make the docker image big and slow down the set up time.
 
 ## Hardware Dependencies
-1. To support claim 1 and run experiments in Section 7.3 and 7.4, an Intel CPU will work;
-2. To support claim 2 and run experiments in Section 7.2, our original experiments use a 48-core Intel Xeon Gold 6248R CPU at 3.00GHz and an NVIDIA RTX A6000 GPU. But to support claim 2, we believe a multi-core Intel CPU and an NVIDIA GPU will work.
+1. To run experiments 1 and 2, an Intel CPU will work;
+2. To run experiment 3, our original experiment uses a 48-core Intel Xeon Gold 6248R CPU at 3.00GHz, while we belive a multi-core Intel CPU should work;
+3. To run experiment 4, our original experiment uses an NVIDIA RTX A6000 GPU, while we believe an NVIDIA GPU in general will work.
 
 ## Getting Started
+### For Experiments 1, 2, 3
+We first pull a docker image from dockerhub:  
+`docker login -u <your-dockerhub-username>`
+`docker pull sibylau/mlir-llvm:omp`
+It may take ~5mins to download the docker image.
+Then we run a container from this docker image:
+`docker run --rm -it --entrypoint bash <image-name-or-id>`
+git clone zenodo UniSparse
+source the bash file under the UniSparse project directory path:  
+`$cd UniSparse && source script/build.sh`
+
+### For Experiment 4
+We do not use the docker image. Please directly clone the zenedo repo:
 
 
 ## Step-by-Step Instructions
+### Experiment 1
+
+
+### Experiment 2
 
+### Experiment 3
 
+### Experiment 4
 
 ## Reusability Guide
 

evaluation/FormatConversion/results/coo_to_cisr.log

@@ -0,0 +1,300 @@
+roadNet-PA_row_major.mtx
+Size of the matrix: 1090920 1090920 1541898
+
+Read from file done, time = 0.729361(s)
+
+Sum done, time = 0.003378(s)
+
+Schedule done, time = 0.057282(s)
+
+Move done, time = 0.040737(s)
+
+Fuse done, time = 0.006423(s)
+
+Fuse done, time = 0.013886(s)
+0.121751
+wikipedia-20051105_row_major.mtx
+Size of the matrix: 1634989 1634989 19753078
+
+Read from file done, time = 7.67207(s)
+
+Sum done, time = 0.02387(s)
+
+Schedule done, time = 0.204234(s)
+
+Move done, time = 0.613762(s)
+
+Fuse done, time = 0.084371(s)
+
+Fuse done, time = 0.101071(s)
+1.02738
+mouse_gene_45K_29M_csr_float32_row_major.mtx
+Size of the matrix: 45101 45101 14506195
+
+Read from file done, time = 4.66598(s)
+
+Sum done, time = 0.02053(s)
+
+Schedule done, time = 0.078583(s)
+
+Move done, time = 0.271226(s)
+
+Fuse done, time = 0.060781(s)
+
+Fuse done, time = 0.064089(s)
+0.495284
+gplus_108K_13M_csr_float32_row_major.mtx
+Size of the matrix: 107614 107614 13673453
+
+Read from file done, time = 4.46023(s)
+
+Sum done, time = 0.019199(s)
+
+Schedule done, time = 0.078054(s)
+
+Move done, time = 0.259289(s)
+
+Fuse done, time = 0.059894(s)
+
+Fuse done, time = 0.061876(s)
+0.478395
+pokec_1633K_31M_csr_float32_row_major.mtx
+Size of the matrix: 1632803 1632803 30622564
+
+Read from file done, time = 10.6273(s)
+
+Sum done, time = 0.036899(s)
+
+Schedule done, time = 0.267732(s)
+
+Move done, time = 0.9401(s)
+
+Fuse done, time = 0.128151(s)
+
+Fuse done, time = 0.15218(s)
+1.52518
+hollywood_1M_113M_row_major.mtx
+Size of the matrix: 1069126 1069126 112613306
+
+Read from file done, time = 42.26(s)
+
+Sum done, time = 0.153701(s)
+
+Schedule done, time = 0.651654(s)
+
+Move done, time = 2.328(s)
+
+Fuse done, time = 0.474178(s)
+
+Fuse done, time = 0.52635(s)
+4.13403
+ogbl_ppa_576K_42M_csr_float32_row_major.mtx
+Size of the matrix: 576289 576289 42463861
+
+Read from file done, time = 14.5706(s)
+
+Sum done, time = 0.056255(s)
+
+Schedule done, time = 0.261021(s)
+
+Move done, time = 0.977207(s)
+
+Fuse done, time = 0.180187(s)
+
+Fuse done, time = 0.206182(s)
+1.68093
+live_journal_5M_69M_csr_float32_row_major.mtx
+Size of the matrix: 4847571 4847571 68475391
+
+Read from file done, time = 23.4763(s)
+
+Sum done, time = 0.083389(s)
+
+Schedule done, time = 0.646846(s)
+
+Move done, time = 2.14508(s)
+
+Fuse done, time = 0.287428(s)
+
+Fuse done, time = 0.356758(s)
+3.51962
+wiki-Vote_row_major.mtx
+Size of the matrix: 8297 8297 103689
+
+Read from file done, time = 0.052644(s)
+
+Sum done, time = 0.00021(s)
+
+Schedule done, time = 0.001081(s)
+
+Move done, time = 0.001899(s)
+
+Fuse done, time = 0.000588(s)
+
+Fuse done, time = 0.000657(s)
+0.00446987
+email-Eu-core_row_major.mtx
+Size of the matrix: 1005 1005 25571
+
+Read from file done, time = 0.017535(s)
+
+Sum done, time = 0.000104(s)
+
+Schedule done, time = 0.000174(s)
+
+Move done, time = 0.000424(s)
+
+Fuse done, time = 0.000157(s)
+
+Fuse done, time = 0.000164(s)
+0.00105906
+nemeth21_row_major.mtx
+Size of the matrix: 9506 9506 591626
+
+Read from file done, time = 0.326929(s)
+
+Sum done, time = 0.000908(s)
+
+Schedule done, time = 0.0025(s)
+
+Move done, time = 0.009191(s)
+
+Fuse done, time = 0.002565(s)
+
+Fuse done, time = 0.0027(s)
+0.017885
+crystm02_row_major.mtx
+Size of the matrix: 13965 13965 168435
+
+Read from file done, time = 0.105582(s)
+
+Sum done, time = 0.000329(s)
+
+Schedule done, time = 0.001547(s)
+
+Move done, time = 0.002845(s)
+
+Fuse done, time = 0.000936(s)
+
+Fuse done, time = 0.001025(s)
+0.00670791
+cant_row_major.mtx
+Size of the matrix: 62451 62451 2034917
+
+Read from file done, time = 1.01595(s)
+
+Sum done, time = 0.003124(s)
+
+Schedule done, time = 0.012889(s)
+
+Move done, time = 0.040015(s)
+
+Fuse done, time = 0.008536(s)
+
+Fuse done, time = 0.009616(s)
+0.074214
+ML_Laplace_row_major.mtx
+Size of the matrix: 377002 377002 27689972
+
+Read from file done, time = 12.4771(s)
+
+Sum done, time = 0.042141(s)
+
+Schedule done, time = 0.166592(s)
+
+Move done, time = 0.662296(s)
+
+Fuse done, time = 0.117457(s)
+
+Fuse done, time = 0.125776(s)
+1.11437
+Transport_row_major.mtx
+Size of the matrix: 1602111 1602111 23500731
+
+Read from file done, time = 11.0639(s)
+
+Sum done, time = 0.036945(s)
+
+Schedule done, time = 0.207547(s)
+
+Move done, time = 0.848522(s)
+
+Fuse done, time = 0.103247(s)
+
+Fuse done, time = 0.106751(s)
+1.30312
+TSOPF_RS_b2383_row_major.mtx
+Size of the matrix: 38120 38120 16171169
+
+Read from file done, time = 7.13491(s)
+
+Sum done, time = 0.024517(s)
+
+Schedule done, time = 0.085628(s)
+
+Move done, time = 0.270983(s)
+
+Fuse done, time = 0.069645(s)
+
+Fuse done, time = 0.072007(s)
+0.522833
+chem_master1_row_major.mtx
+Size of the matrix: 40401 40401 201201
+
+Read from file done, time = 0.119561(s)
+
+Sum done, time = 0.00039(s)
+
+Schedule done, time = 0.002867(s)
+
+Move done, time = 0.003933(s)
+
+Fuse done, time = 0.001109(s)
+
+Fuse done, time = 0.001207(s)
+0.00954485
+majorbasis_row_major.mtx
+Size of the matrix: 160000 160000 1750416
+
+Read from file done, time = 0.839045(s)
+
+Sum done, time = 0.002756(s)
+
+Schedule done, time = 0.014406(s)
+
+Move done, time = 0.036031(s)
+
+Fuse done, time = 0.007406(s)
+
+Fuse done, time = 0.007806(s)
+0.0684419
+shyy161_row_major.mtx
+Size of the matrix: 76480 76480 329762
+
+Read from file done, time = 0.153457(s)
+
+Sum done, time = 0.000531(s)
+
+Schedule done, time = 0.004147(s)
+
+Move done, time = 0.005855(s)
+
+Fuse done, time = 0.001438(s)
+
+Fuse done, time = 0.001689(s)
+0.0136809
+Baumann_row_major.mtx
+Size of the matrix: 112211 112211 760631
+
+Read from file done, time = 0.420328(s)
+
+Sum done, time = 0.001153(s)
+
+Schedule done, time = 0.008881(s)
+
+Move done, time = 0.014495(s)
+
+Fuse done, time = 0.0033(s)
+
+Fuse done, time = 0.003596(s)
+0.0314491

evaluation/FormatConversion/results/coo_to_dia.log

@@ -0,0 +1,26 @@
+chem_master1_row_major.mtx
+Size of the matrix: 40401 40401 201201
+Need to have 5 vectors
+0.00416493
+Enter the check function
+Check Success
+majorbasis_row_major.mtx
+Size of the matrix: 160000 160000 1750416
+Need to have 22 vectors
+0.0351272
+Enter the check function
+Check Success
+shyy161_row_major.mtx
+Size of the matrix: 76480 76480 329762
+Need to have 7 vectors
+0.00670195
+Enter the check function
+Check Success
+Baumann_row_major.mtx
+Size of the matrix: 112211 112211 760631
+Need to have 7 vectors
+0.0129969
+Enter the check function
+Check Failed
+coo_to_dia_4: /work/shared/users/staff/zz546/Sparse_Layout_Dialect/lib/Runtime/SparlayUtils.cpp:2092: void _mlir_ciface_sptCheck(void*, void*): Assertion `0' failed.
+Aborted (core dumped)