[GH-PAGES] Updated website
This commit is contained in:
Philippe Tillet
@@ -1,4 +1,4 @@
|
||||
# Sphinx build info version 1
|
||||
# This file hashes the configuration used when building these files. When it is not found, a full rebuild will be done.
|
||||
config: f293410811a23e0582da8ef1fa35bfc6
|
||||
config: 4d6b85d7c9a7d93ae015e5e1fbc4850d
|
||||
tags: 645f666f9bcd5a90fca523b33c5a78b7
|
||||
|
||||
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
|
Before Width: | Height: | Size: 23 KiB After Width: | Height: | Size: 24 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 15 KiB After Width: | Height: | Size: 16 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 36 KiB After Width: | Height: | Size: 36 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 23 KiB After Width: | Height: | Size: 23 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 60 KiB After Width: | Height: | Size: 58 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 34 KiB After Width: | Height: | Size: 33 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 35 KiB After Width: | Height: | Size: 35 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 22 KiB After Width: | Height: | Size: 22 KiB |
@@ -234,8 +234,8 @@ We can now run the decorated function above. Pass `print_data=True` to see the p
|
||||
size Triton Torch
|
||||
0 4096.0 9.600000 9.600000
|
||||
1 8192.0 19.200000 19.200000
|
||||
2 16384.0 38.400001 38.400001
|
||||
3 32768.0 63.999998 63.999998
|
||||
2 16384.0 31.999999 38.400001
|
||||
3 32768.0 76.800002 63.999998
|
||||
4 65536.0 127.999995 127.999995
|
||||
5 131072.0 219.428568 219.428568
|
||||
6 262144.0 384.000001 384.000001
|
||||
@@ -245,7 +245,7 @@ We can now run the decorated function above. Pass `print_data=True` to see the p
|
||||
10 4194304.0 780.190482 780.190482
|
||||
11 8388608.0 812.429770 812.429770
|
||||
12 16777216.0 833.084721 833.084721
|
||||
13 33554432.0 842.004273 843.811163
|
||||
13 33554432.0 842.004273 842.906750
|
||||
14 67108864.0 847.448255 848.362445
|
||||
15 134217728.0 849.737435 850.656574
|
||||
|
||||
@@ -255,7 +255,7 @@ We can now run the decorated function above. Pass `print_data=True` to see the p
|
||||
|
||||
.. rst-class:: sphx-glr-timing
|
||||
|
||||
**Total running time of the script:** ( 1 minutes 50.498 seconds)
|
||||
**Total running time of the script:** ( 1 minutes 50.715 seconds)
|
||||
|
||||
|
||||
.. _sphx_glr_download_getting-started_tutorials_01-vector-add.py:
|
||||
|
||||
@@ -278,17 +278,17 @@ We will then compare its performance against (1) :code:`torch.softmax` and (2) t
|
||||
|
||||
softmax-performance:
|
||||
N Triton Torch (native) Torch (jit)
|
||||
0 256.0 546.133347 546.133347 190.511628
|
||||
1 384.0 614.400016 585.142862 153.600004
|
||||
2 512.0 655.360017 606.814814 154.566038
|
||||
0 256.0 546.133347 546.133347 188.321838
|
||||
1 384.0 614.400016 585.142862 151.703707
|
||||
2 512.0 655.360017 606.814814 156.038096
|
||||
3 640.0 706.206879 640.000002 160.000000
|
||||
4 768.0 722.823517 664.216187 162.754967
|
||||
.. ... ... ... ...
|
||||
93 12160.0 812.359066 406.179533 198.530610
|
||||
94 12288.0 812.429770 415.222812 198.794749
|
||||
95 12416.0 812.498981 412.149375 198.556711
|
||||
96 12544.0 810.925276 412.546756 198.716830
|
||||
97 12672.0 811.007961 412.097543 198.776477
|
||||
93 12160.0 812.359066 405.755985 198.834951
|
||||
94 12288.0 812.429770 415.222812 199.096718
|
||||
95 12416.0 812.498981 411.722274 198.655991
|
||||
96 12544.0 810.925276 412.971190 198.913776
|
||||
97 12672.0 811.007961 412.516771 199.069228
|
||||
|
||||
[98 rows x 4 columns]
|
||||
|
||||
@@ -306,7 +306,7 @@ In the above plot, we can see that:
|
||||
|
||||
.. rst-class:: sphx-glr-timing
|
||||
|
||||
**Total running time of the script:** ( 3 minutes 32.107 seconds)
|
||||
**Total running time of the script:** ( 3 minutes 30.914 seconds)
|
||||
|
||||
|
||||
.. _sphx_glr_download_getting-started_tutorials_02-fused-softmax.py:
|
||||
|
||||
@@ -459,37 +459,37 @@ We can now compare the performance of our kernel against that of cuBLAS. Here we
|
||||
|
||||
matmul-performance:
|
||||
M cuBLAS ... Triton Triton (+ LeakyReLU)
|
||||
0 256.0 2.978909 ... 2.978909 3.276800
|
||||
0 256.0 2.730667 ... 2.978909 2.978909
|
||||
1 384.0 7.372800 ... 8.507077 8.507077
|
||||
2 512.0 14.563555 ... 16.384000 16.384000
|
||||
3 640.0 22.260869 ... 24.380953 24.380953
|
||||
4 768.0 32.768000 ... 35.389441 34.028308
|
||||
5 896.0 39.025776 ... 40.140799 39.025776
|
||||
6 1024.0 49.932191 ... 53.773130 52.428801
|
||||
7 1152.0 44.566925 ... 47.396572 47.396572
|
||||
7 1152.0 45.242181 ... 48.161033 47.396572
|
||||
8 1280.0 51.200001 ... 57.690139 57.690139
|
||||
9 1408.0 64.138541 ... 68.147202 67.305878
|
||||
10 1536.0 80.430545 ... 81.355034 79.526831
|
||||
9 1408.0 64.138541 ... 69.009825 67.305878
|
||||
10 1536.0 80.430545 ... 80.430545 79.526831
|
||||
11 1664.0 62.929456 ... 63.372618 62.492442
|
||||
12 1792.0 72.512412 ... 73.460287 59.467852
|
||||
13 1920.0 69.120002 ... 71.626943 71.257735
|
||||
14 2048.0 73.908442 ... 78.398206 77.314362
|
||||
15 2176.0 83.500614 ... 87.876193 86.367588
|
||||
16 2304.0 68.446623 ... 78.064941 77.307030
|
||||
17 2432.0 71.305746 ... 86.711310 85.653855
|
||||
18 2560.0 78.019048 ... 82.956960 81.715711
|
||||
19 2688.0 83.922689 ... 90.316801 90.102270
|
||||
20 2816.0 79.587973 ... 83.873477 83.873477
|
||||
21 2944.0 81.832567 ... 83.617504 82.921853
|
||||
22 3072.0 81.707223 ... 90.164177 88.197981
|
||||
23 3200.0 82.262212 ... 97.190583 95.238096
|
||||
24 3328.0 82.843841 ... 85.196803 84.200347
|
||||
25 3456.0 79.430113 ... 85.133652 82.604067
|
||||
26 3584.0 87.381330 ... 99.244365 91.563533
|
||||
27 3712.0 84.874549 ... 82.902362 86.044224
|
||||
28 3840.0 84.550462 ... 93.405401 84.809814
|
||||
29 3968.0 92.512459 ... 84.975809 90.859224
|
||||
30 4096.0 86.313653 ... 93.792965 90.321484
|
||||
16 2304.0 68.251065 ... 78.064941 77.307030
|
||||
17 2432.0 71.305746 ... 86.179335 85.393507
|
||||
18 2560.0 77.833728 ... 82.539044 81.512437
|
||||
19 2688.0 83.737433 ... 91.185232 89.254248
|
||||
20 2816.0 79.879498 ... 82.602666 83.392363
|
||||
21 2944.0 82.102191 ... 82.990890 83.337844
|
||||
22 3072.0 80.202695 ... 89.170242 87.381335
|
||||
23 3200.0 82.474230 ... 96.676741 95.238096
|
||||
24 3328.0 82.843841 ... 86.062515 84.795401
|
||||
25 3456.0 81.026701 ... 91.200871 87.347312
|
||||
26 3584.0 87.381330 ... 95.350361 98.268190
|
||||
27 3712.0 85.970176 ... 89.353616 87.552452
|
||||
28 3840.0 79.192264 ... 91.853823 85.796739
|
||||
29 3968.0 87.850207 ... 86.449828 89.988156
|
||||
30 4096.0 86.509232 ... 92.948562 87.352901
|
||||
|
||||
[31 rows x 5 columns]
|
||||
|
||||
@@ -499,7 +499,7 @@ We can now compare the performance of our kernel against that of cuBLAS. Here we
|
||||
|
||||
.. rst-class:: sphx-glr-timing
|
||||
|
||||
**Total running time of the script:** ( 6 minutes 38.502 seconds)
|
||||
**Total running time of the script:** ( 7 minutes 14.457 seconds)
|
||||
|
||||
|
||||
.. _sphx_glr_download_getting-started_tutorials_03-matrix-multiplication.py:
|
||||
|
||||
@@ -240,7 +240,7 @@ References
|
||||
|
||||
.. rst-class:: sphx-glr-timing
|
||||
|
||||
**Total running time of the script:** ( 0 minutes 0.012 seconds)
|
||||
**Total running time of the script:** ( 0 minutes 0.282 seconds)
|
||||
|
||||
|
||||
.. _sphx_glr_download_getting-started_tutorials_04-low-memory-dropout.py:
|
||||
|
||||
@@ -38,35 +38,35 @@ Layer Normalization
|
||||
|
||||
layer-norm:
|
||||
N Triton Torch Apex
|
||||
0 1024.0 606.814814 277.694907 468.114273
|
||||
0 1024.0 585.142849 277.694907 468.114273
|
||||
1 1536.0 630.153868 323.368435 511.999982
|
||||
2 2048.0 682.666643 334.367358 520.126988
|
||||
3 2560.0 694.237267 365.714281 512.000013
|
||||
4 3072.0 712.347810 378.092307 496.484863
|
||||
5 3584.0 725.873439 384.859062 451.527536
|
||||
6 4096.0 728.177767 381.023256 455.111095
|
||||
7 4608.0 670.254540 394.267384 426.173427
|
||||
8 5120.0 688.403381 397.669909 422.268057
|
||||
9 5632.0 704.000002 395.228063 415.262685
|
||||
2 2048.0 682.666643 337.814445 520.126988
|
||||
3 2560.0 694.237267 365.714281 518.481028
|
||||
4 3072.0 712.347810 375.206126 496.484863
|
||||
5 3584.0 725.873439 384.859062 455.111115
|
||||
6 4096.0 728.177767 381.023256 442.810792
|
||||
7 4608.0 670.254540 396.387087 426.173427
|
||||
8 5120.0 688.403381 397.669909 426.666652
|
||||
9 5632.0 698.542675 398.725657 411.470331
|
||||
10 6144.0 697.191505 402.885254 409.600010
|
||||
11 6656.0 700.631610 400.360920 400.360920
|
||||
12 7168.0 690.891575 396.844306 387.459443
|
||||
13 7680.0 678.895043 393.846167 386.415087
|
||||
14 8192.0 636.271854 393.609605 371.308771
|
||||
15 8704.0 627.315309 389.005597 381.545190
|
||||
16 9216.0 606.814809 407.337026 383.999986
|
||||
17 9728.0 587.350922 409.599987 383.369452
|
||||
18 10240.0 564.965524 408.578556 382.803739
|
||||
19 10752.0 547.872604 411.559798 381.445676
|
||||
20 11264.0 533.207081 406.826188 373.134567
|
||||
21 11776.0 520.486200 409.599991 377.587162
|
||||
22 12288.0 514.680630 413.911572 383.251457
|
||||
23 12800.0 504.433489 410.420828 376.470582
|
||||
24 13312.0 494.180982 405.699062 376.976995
|
||||
25 13824.0 482.934503 411.888257 379.389355
|
||||
26 14336.0 471.967074 406.695045 374.185964
|
||||
27 14848.0 461.297068 408.192434 375.304904
|
||||
28 15360.0 454.269882 406.214870 378.092307
|
||||
11 6656.0 700.631610 400.360920 398.861429
|
||||
12 7168.0 690.891575 382.293315 382.293315
|
||||
13 7680.0 678.895043 392.587863 386.415087
|
||||
14 8192.0 636.271854 392.431125 374.491442
|
||||
15 8704.0 624.502255 392.292962 380.502740
|
||||
16 9216.0 606.814809 403.989025 383.002605
|
||||
17 9728.0 587.350922 407.455499 382.427505
|
||||
18 10240.0 566.920437 407.562184 381.911416
|
||||
19 10752.0 547.872604 410.577576 380.601764
|
||||
20 11264.0 533.207081 396.096702 369.311483
|
||||
21 11776.0 521.927959 407.826843 377.587162
|
||||
22 12288.0 516.031509 413.042029 382.505826
|
||||
23 12800.0 504.433489 408.782457 376.470582
|
||||
24 13312.0 494.180982 401.871683 375.647260
|
||||
25 13824.0 482.934503 409.600016 378.092325
|
||||
26 14336.0 471.967074 398.914774 372.969090
|
||||
27 14848.0 461.297068 403.341254 374.712936
|
||||
28 15360.0 454.269882 406.887417 378.092307
|
||||
29 15872.0 447.887117 406.974373 376.225175
|
||||
|
||||
|
||||
@@ -393,7 +393,7 @@ Layer Normalization
|
||||
|
||||
.. rst-class:: sphx-glr-timing
|
||||
|
||||
**Total running time of the script:** ( 5 minutes 38.270 seconds)
|
||||
**Total running time of the script:** ( 5 minutes 38.714 seconds)
|
||||
|
||||
|
||||
.. _sphx_glr_download_getting-started_tutorials_05-layer-norm.py:
|
||||
|
||||
@@ -385,7 +385,7 @@ This is a Triton implementation of the Flash Attention algorithm
|
||||
|
||||
.. rst-class:: sphx-glr-timing
|
||||
|
||||
**Total running time of the script:** ( 0 minutes 0.073 seconds)
|
||||
**Total running time of the script:** ( 0 minutes 0.074 seconds)
|
||||
|
||||
|
||||
.. _sphx_glr_download_getting-started_tutorials_06-fused-attention.py:
|
||||
|
||||
@@ -152,7 +152,7 @@ We can also customize the libdevice library path by passing the path to the `lib
|
||||
|
||||
.. rst-class:: sphx-glr-timing
|
||||
|
||||
**Total running time of the script:** ( 0 minutes 0.010 seconds)
|
||||
**Total running time of the script:** ( 0 minutes 0.250 seconds)
|
||||
|
||||
|
||||
.. _sphx_glr_download_getting-started_tutorials_07-libdevice-function.py:
|
||||
|
||||
@@ -5,20 +5,20 @@
|
||||
|
||||
Computation times
|
||||
=================
|
||||
**17:39.472** total execution time for **getting-started_tutorials** files:
|
||||
**18:15.408** total execution time for **getting-started_tutorials** files:
|
||||
|
||||
+---------------------------------------------------------------------------------------------------------+-----------+--------+
|
||||
| :ref:`sphx_glr_getting-started_tutorials_03-matrix-multiplication.py` (``03-matrix-multiplication.py``) | 06:38.502 | 0.0 MB |
|
||||
| :ref:`sphx_glr_getting-started_tutorials_03-matrix-multiplication.py` (``03-matrix-multiplication.py``) | 07:14.457 | 0.0 MB |
|
||||
+---------------------------------------------------------------------------------------------------------+-----------+--------+
|
||||
| :ref:`sphx_glr_getting-started_tutorials_05-layer-norm.py` (``05-layer-norm.py``) | 05:38.270 | 0.0 MB |
|
||||
| :ref:`sphx_glr_getting-started_tutorials_05-layer-norm.py` (``05-layer-norm.py``) | 05:38.714 | 0.0 MB |
|
||||
+---------------------------------------------------------------------------------------------------------+-----------+--------+
|
||||
| :ref:`sphx_glr_getting-started_tutorials_02-fused-softmax.py` (``02-fused-softmax.py``) | 03:32.107 | 0.0 MB |
|
||||
| :ref:`sphx_glr_getting-started_tutorials_02-fused-softmax.py` (``02-fused-softmax.py``) | 03:30.914 | 0.0 MB |
|
||||
+---------------------------------------------------------------------------------------------------------+-----------+--------+
|
||||
| :ref:`sphx_glr_getting-started_tutorials_01-vector-add.py` (``01-vector-add.py``) | 01:50.498 | 0.0 MB |
|
||||
| :ref:`sphx_glr_getting-started_tutorials_01-vector-add.py` (``01-vector-add.py``) | 01:50.715 | 0.0 MB |
|
||||
+---------------------------------------------------------------------------------------------------------+-----------+--------+
|
||||
| :ref:`sphx_glr_getting-started_tutorials_06-fused-attention.py` (``06-fused-attention.py``) | 00:00.073 | 0.0 MB |
|
||||
| :ref:`sphx_glr_getting-started_tutorials_04-low-memory-dropout.py` (``04-low-memory-dropout.py``) | 00:00.282 | 0.0 MB |
|
||||
+---------------------------------------------------------------------------------------------------------+-----------+--------+
|
||||
| :ref:`sphx_glr_getting-started_tutorials_04-low-memory-dropout.py` (``04-low-memory-dropout.py``) | 00:00.012 | 0.0 MB |
|
||||
| :ref:`sphx_glr_getting-started_tutorials_07-libdevice-function.py` (``07-libdevice-function.py``) | 00:00.250 | 0.0 MB |
|
||||
+---------------------------------------------------------------------------------------------------------+-----------+--------+
|
||||
| :ref:`sphx_glr_getting-started_tutorials_07-libdevice-function.py` (``07-libdevice-function.py``) | 00:00.010 | 0.0 MB |
|
||||
| :ref:`sphx_glr_getting-started_tutorials_06-fused-attention.py` (``06-fused-attention.py``) | 00:00.074 | 0.0 MB |
|
||||
+---------------------------------------------------------------------------------------------------------+-----------+--------+
|
||||
|
||||
@@ -326,8 +326,8 @@ for different problem sizes.</p>
|
||||
size Triton Torch
|
||||
0 4096.0 9.600000 9.600000
|
||||
1 8192.0 19.200000 19.200000
|
||||
2 16384.0 38.400001 38.400001
|
||||
3 32768.0 63.999998 63.999998
|
||||
2 16384.0 31.999999 38.400001
|
||||
3 32768.0 76.800002 63.999998
|
||||
4 65536.0 127.999995 127.999995
|
||||
5 131072.0 219.428568 219.428568
|
||||
6 262144.0 384.000001 384.000001
|
||||
@@ -337,12 +337,12 @@ for different problem sizes.</p>
|
||||
10 4194304.0 780.190482 780.190482
|
||||
11 8388608.0 812.429770 812.429770
|
||||
12 16777216.0 833.084721 833.084721
|
||||
13 33554432.0 842.004273 843.811163
|
||||
13 33554432.0 842.004273 842.906750
|
||||
14 67108864.0 847.448255 848.362445
|
||||
15 134217728.0 849.737435 850.656574
|
||||
</pre></div>
|
||||
</div>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 50.498 seconds)</p>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 50.715 seconds)</p>
|
||||
<div class="sphx-glr-footer class sphx-glr-footer-example docutils container" id="sphx-glr-download-getting-started-tutorials-01-vector-add-py">
|
||||
<div class="sphx-glr-download sphx-glr-download-python docutils container">
|
||||
<p><a class="reference download internal" download="" href="../../_downloads/62d97d49a32414049819dd8bb8378080/01-vector-add.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">01-vector-add.py</span></code></a></p>
|
||||
|
||||
@@ -371,17 +371,17 @@ We will then compare its performance against (1) <code class="code docutils lite
|
||||
<p class="sphx-glr-script-out">Out:</p>
|
||||
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>softmax-performance:
|
||||
N Triton Torch (native) Torch (jit)
|
||||
0 256.0 546.133347 546.133347 190.511628
|
||||
1 384.0 614.400016 585.142862 153.600004
|
||||
2 512.0 655.360017 606.814814 154.566038
|
||||
0 256.0 546.133347 546.133347 188.321838
|
||||
1 384.0 614.400016 585.142862 151.703707
|
||||
2 512.0 655.360017 606.814814 156.038096
|
||||
3 640.0 706.206879 640.000002 160.000000
|
||||
4 768.0 722.823517 664.216187 162.754967
|
||||
.. ... ... ... ...
|
||||
93 12160.0 812.359066 406.179533 198.530610
|
||||
94 12288.0 812.429770 415.222812 198.794749
|
||||
95 12416.0 812.498981 412.149375 198.556711
|
||||
96 12544.0 810.925276 412.546756 198.716830
|
||||
97 12672.0 811.007961 412.097543 198.776477
|
||||
93 12160.0 812.359066 405.755985 198.834951
|
||||
94 12288.0 812.429770 415.222812 199.096718
|
||||
95 12416.0 812.498981 411.722274 198.655991
|
||||
96 12544.0 810.925276 412.971190 198.913776
|
||||
97 12672.0 811.007961 412.516771 199.069228
|
||||
|
||||
[98 rows x 4 columns]
|
||||
</pre></div>
|
||||
@@ -394,7 +394,7 @@ We will then compare its performance against (1) <code class="code docutils lite
|
||||
Note however that the PyTorch <cite>softmax</cite> operation is more general and will works on tensors of any shape.</p></li>
|
||||
</ul>
|
||||
</div></blockquote>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 3 minutes 32.107 seconds)</p>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 3 minutes 30.914 seconds)</p>
|
||||
<div class="sphx-glr-footer class sphx-glr-footer-example docutils container" id="sphx-glr-download-getting-started-tutorials-02-fused-softmax-py">
|
||||
<div class="sphx-glr-download sphx-glr-download-python docutils container">
|
||||
<p><a class="reference download internal" download="" href="../../_downloads/d91442ac2982c4e0cc3ab0f43534afbc/02-fused-softmax.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">02-fused-softmax.py</span></code></a></p>
|
||||
|
||||
@@ -567,42 +567,42 @@ torch_output=tensor([[ 1.1045, -36.9688, 31.4688, ..., -11.3906, 24.4531, -3
|
||||
<p class="sphx-glr-script-out">Out:</p>
|
||||
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>matmul-performance:
|
||||
M cuBLAS ... Triton Triton (+ LeakyReLU)
|
||||
0 256.0 2.978909 ... 2.978909 3.276800
|
||||
0 256.0 2.730667 ... 2.978909 2.978909
|
||||
1 384.0 7.372800 ... 8.507077 8.507077
|
||||
2 512.0 14.563555 ... 16.384000 16.384000
|
||||
3 640.0 22.260869 ... 24.380953 24.380953
|
||||
4 768.0 32.768000 ... 35.389441 34.028308
|
||||
5 896.0 39.025776 ... 40.140799 39.025776
|
||||
6 1024.0 49.932191 ... 53.773130 52.428801
|
||||
7 1152.0 44.566925 ... 47.396572 47.396572
|
||||
7 1152.0 45.242181 ... 48.161033 47.396572
|
||||
8 1280.0 51.200001 ... 57.690139 57.690139
|
||||
9 1408.0 64.138541 ... 68.147202 67.305878
|
||||
10 1536.0 80.430545 ... 81.355034 79.526831
|
||||
9 1408.0 64.138541 ... 69.009825 67.305878
|
||||
10 1536.0 80.430545 ... 80.430545 79.526831
|
||||
11 1664.0 62.929456 ... 63.372618 62.492442
|
||||
12 1792.0 72.512412 ... 73.460287 59.467852
|
||||
13 1920.0 69.120002 ... 71.626943 71.257735
|
||||
14 2048.0 73.908442 ... 78.398206 77.314362
|
||||
15 2176.0 83.500614 ... 87.876193 86.367588
|
||||
16 2304.0 68.446623 ... 78.064941 77.307030
|
||||
17 2432.0 71.305746 ... 86.711310 85.653855
|
||||
18 2560.0 78.019048 ... 82.956960 81.715711
|
||||
19 2688.0 83.922689 ... 90.316801 90.102270
|
||||
20 2816.0 79.587973 ... 83.873477 83.873477
|
||||
21 2944.0 81.832567 ... 83.617504 82.921853
|
||||
22 3072.0 81.707223 ... 90.164177 88.197981
|
||||
23 3200.0 82.262212 ... 97.190583 95.238096
|
||||
24 3328.0 82.843841 ... 85.196803 84.200347
|
||||
25 3456.0 79.430113 ... 85.133652 82.604067
|
||||
26 3584.0 87.381330 ... 99.244365 91.563533
|
||||
27 3712.0 84.874549 ... 82.902362 86.044224
|
||||
28 3840.0 84.550462 ... 93.405401 84.809814
|
||||
29 3968.0 92.512459 ... 84.975809 90.859224
|
||||
30 4096.0 86.313653 ... 93.792965 90.321484
|
||||
16 2304.0 68.251065 ... 78.064941 77.307030
|
||||
17 2432.0 71.305746 ... 86.179335 85.393507
|
||||
18 2560.0 77.833728 ... 82.539044 81.512437
|
||||
19 2688.0 83.737433 ... 91.185232 89.254248
|
||||
20 2816.0 79.879498 ... 82.602666 83.392363
|
||||
21 2944.0 82.102191 ... 82.990890 83.337844
|
||||
22 3072.0 80.202695 ... 89.170242 87.381335
|
||||
23 3200.0 82.474230 ... 96.676741 95.238096
|
||||
24 3328.0 82.843841 ... 86.062515 84.795401
|
||||
25 3456.0 81.026701 ... 91.200871 87.347312
|
||||
26 3584.0 87.381330 ... 95.350361 98.268190
|
||||
27 3712.0 85.970176 ... 89.353616 87.552452
|
||||
28 3840.0 79.192264 ... 91.853823 85.796739
|
||||
29 3968.0 87.850207 ... 86.449828 89.988156
|
||||
30 4096.0 86.509232 ... 92.948562 87.352901
|
||||
|
||||
[31 rows x 5 columns]
|
||||
</pre></div>
|
||||
</div>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 6 minutes 38.502 seconds)</p>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 7 minutes 14.457 seconds)</p>
|
||||
<div class="sphx-glr-footer class sphx-glr-footer-example docutils container" id="sphx-glr-download-getting-started-tutorials-03-matrix-multiplication-py">
|
||||
<div class="sphx-glr-download sphx-glr-download-python docutils container">
|
||||
<p><a class="reference download internal" download="" href="../../_downloads/d5fee5b55a64e47f1b5724ec39adf171/03-matrix-multiplication.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">03-matrix-multiplication.py</span></code></a></p>
|
||||
|
||||
@@ -374,7 +374,7 @@ to explore the <cite>triton/language/random</cite> folder!</p>
|
||||
<dd><p>Nitish Srivastava and Geoffrey Hinton and Alex Krizhevsky and Ilya Sutskever and Ruslan Salakhutdinov, “Dropout: A Simple Way to Prevent Neural Networks from Overfitting”, JMLR 2014</p>
|
||||
</dd>
|
||||
</dl>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 0 minutes 0.012 seconds)</p>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 0 minutes 0.282 seconds)</p>
|
||||
<div class="sphx-glr-footer class sphx-glr-footer-example docutils container" id="sphx-glr-download-getting-started-tutorials-04-low-memory-dropout-py">
|
||||
<div class="sphx-glr-download sphx-glr-download-python docutils container">
|
||||
<p><a class="reference download internal" download="" href="../../_downloads/c9aed78977a4c05741d675a38dde3d7d/04-low-memory-dropout.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">04-low-memory-dropout.py</span></code></a></p>
|
||||
|
||||
@@ -196,35 +196,35 @@ to download the full example code</p>
|
||||
<p class="sphx-glr-script-out">Out:</p>
|
||||
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>layer-norm:
|
||||
N Triton Torch Apex
|
||||
0 1024.0 606.814814 277.694907 468.114273
|
||||
0 1024.0 585.142849 277.694907 468.114273
|
||||
1 1536.0 630.153868 323.368435 511.999982
|
||||
2 2048.0 682.666643 334.367358 520.126988
|
||||
3 2560.0 694.237267 365.714281 512.000013
|
||||
4 3072.0 712.347810 378.092307 496.484863
|
||||
5 3584.0 725.873439 384.859062 451.527536
|
||||
6 4096.0 728.177767 381.023256 455.111095
|
||||
7 4608.0 670.254540 394.267384 426.173427
|
||||
8 5120.0 688.403381 397.669909 422.268057
|
||||
9 5632.0 704.000002 395.228063 415.262685
|
||||
2 2048.0 682.666643 337.814445 520.126988
|
||||
3 2560.0 694.237267 365.714281 518.481028
|
||||
4 3072.0 712.347810 375.206126 496.484863
|
||||
5 3584.0 725.873439 384.859062 455.111115
|
||||
6 4096.0 728.177767 381.023256 442.810792
|
||||
7 4608.0 670.254540 396.387087 426.173427
|
||||
8 5120.0 688.403381 397.669909 426.666652
|
||||
9 5632.0 698.542675 398.725657 411.470331
|
||||
10 6144.0 697.191505 402.885254 409.600010
|
||||
11 6656.0 700.631610 400.360920 400.360920
|
||||
12 7168.0 690.891575 396.844306 387.459443
|
||||
13 7680.0 678.895043 393.846167 386.415087
|
||||
14 8192.0 636.271854 393.609605 371.308771
|
||||
15 8704.0 627.315309 389.005597 381.545190
|
||||
16 9216.0 606.814809 407.337026 383.999986
|
||||
17 9728.0 587.350922 409.599987 383.369452
|
||||
18 10240.0 564.965524 408.578556 382.803739
|
||||
19 10752.0 547.872604 411.559798 381.445676
|
||||
20 11264.0 533.207081 406.826188 373.134567
|
||||
21 11776.0 520.486200 409.599991 377.587162
|
||||
22 12288.0 514.680630 413.911572 383.251457
|
||||
23 12800.0 504.433489 410.420828 376.470582
|
||||
24 13312.0 494.180982 405.699062 376.976995
|
||||
25 13824.0 482.934503 411.888257 379.389355
|
||||
26 14336.0 471.967074 406.695045 374.185964
|
||||
27 14848.0 461.297068 408.192434 375.304904
|
||||
28 15360.0 454.269882 406.214870 378.092307
|
||||
11 6656.0 700.631610 400.360920 398.861429
|
||||
12 7168.0 690.891575 382.293315 382.293315
|
||||
13 7680.0 678.895043 392.587863 386.415087
|
||||
14 8192.0 636.271854 392.431125 374.491442
|
||||
15 8704.0 624.502255 392.292962 380.502740
|
||||
16 9216.0 606.814809 403.989025 383.002605
|
||||
17 9728.0 587.350922 407.455499 382.427505
|
||||
18 10240.0 566.920437 407.562184 381.911416
|
||||
19 10752.0 547.872604 410.577576 380.601764
|
||||
20 11264.0 533.207081 396.096702 369.311483
|
||||
21 11776.0 521.927959 407.826843 377.587162
|
||||
22 12288.0 516.031509 413.042029 382.505826
|
||||
23 12800.0 504.433489 408.782457 376.470582
|
||||
24 13312.0 494.180982 401.871683 375.647260
|
||||
25 13824.0 482.934503 409.600016 378.092325
|
||||
26 14336.0 471.967074 398.914774 372.969090
|
||||
27 14848.0 461.297068 403.341254 374.712936
|
||||
28 15360.0 454.269882 406.887417 378.092307
|
||||
29 15872.0 447.887117 406.974373 376.225175
|
||||
</pre></div>
|
||||
</div>
|
||||
@@ -543,7 +543,7 @@ to download the full example code</p>
|
||||
<span class="n">bench_layer_norm</span><span class="o">.</span><span class="n">run</span><span class="p">(</span><span class="n">save_path</span><span class="o">=</span><span class="s1">'.'</span><span class="p">,</span> <span class="n">print_data</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
|
||||
</pre></div>
|
||||
</div>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 5 minutes 38.270 seconds)</p>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 5 minutes 38.714 seconds)</p>
|
||||
<div class="sphx-glr-footer class sphx-glr-footer-example docutils container" id="sphx-glr-download-getting-started-tutorials-05-layer-norm-py">
|
||||
<div class="sphx-glr-download sphx-glr-download-python docutils container">
|
||||
<p><a class="reference download internal" download="" href="../../_downloads/935c0dd0fbeb4b2e69588471cbb2d4b2/05-layer-norm.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">05-layer-norm.py</span></code></a></p>
|
||||
|
||||
@@ -543,7 +543,7 @@ to download the full example code</p>
|
||||
<span class="c1"># bench_flash_attention.run(save_path='.', print_data=True)</span>
|
||||
</pre></div>
|
||||
</div>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 0 minutes 0.073 seconds)</p>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 0 minutes 0.074 seconds)</p>
|
||||
<div class="sphx-glr-footer class sphx-glr-footer-example docutils container" id="sphx-glr-download-getting-started-tutorials-06-fused-attention-py">
|
||||
<div class="sphx-glr-download sphx-glr-download-python docutils container">
|
||||
<p><a class="reference download internal" download="" href="../../_downloads/54a35f6ec55f9746935b9566fb6bb1df/06-fused-attention.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">06-fused-attention.py</span></code></a></p>
|
||||
|
||||
@@ -276,7 +276,7 @@ tensor([0.4105, 0.5430, 0.0249, ..., 0.0424, 0.5351, 0.8149], device='cuda:
|
||||
The maximum difference between torch and triton is 2.384185791015625e-07
|
||||
</pre></div>
|
||||
</div>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 0 minutes 0.010 seconds)</p>
|
||||
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 0 minutes 0.250 seconds)</p>
|
||||
<div class="sphx-glr-footer class sphx-glr-footer-example docutils container" id="sphx-glr-download-getting-started-tutorials-07-libdevice-function-py">
|
||||
<div class="sphx-glr-download sphx-glr-download-python docutils container">
|
||||
<p><a class="reference download internal" download="" href="../../_downloads/3ff29f967ace7985da24aab10352fc76/07-libdevice-function.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">07-libdevice-function.py</span></code></a></p>
|
||||
|
||||
@@ -174,7 +174,7 @@
|
||||
|
||||
<div class="section" id="computation-times">
|
||||
<span id="sphx-glr-getting-started-tutorials-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
|
||||
<p><strong>17:39.472</strong> total execution time for <strong>getting-started_tutorials</strong> files:</p>
|
||||
<p><strong>18:15.408</strong> total execution time for <strong>getting-started_tutorials</strong> files:</p>
|
||||
<table class="docutils align-default">
|
||||
<colgroup>
|
||||
<col style="width: 85%" />
|
||||
@@ -183,31 +183,31 @@
|
||||
</colgroup>
|
||||
<tbody>
|
||||
<tr class="row-odd"><td><p><a class="reference internal" href="03-matrix-multiplication.html#sphx-glr-getting-started-tutorials-03-matrix-multiplication-py"><span class="std std-ref">Matrix Multiplication</span></a> (<code class="docutils literal notranslate"><span class="pre">03-matrix-multiplication.py</span></code>)</p></td>
|
||||
<td><p>06:38.502</p></td>
|
||||
<td><p>07:14.457</p></td>
|
||||
<td><p>0.0 MB</p></td>
|
||||
</tr>
|
||||
<tr class="row-even"><td><p><a class="reference internal" href="05-layer-norm.html#sphx-glr-getting-started-tutorials-05-layer-norm-py"><span class="std std-ref">Layer Normalization</span></a> (<code class="docutils literal notranslate"><span class="pre">05-layer-norm.py</span></code>)</p></td>
|
||||
<td><p>05:38.270</p></td>
|
||||
<td><p>05:38.714</p></td>
|
||||
<td><p>0.0 MB</p></td>
|
||||
</tr>
|
||||
<tr class="row-odd"><td><p><a class="reference internal" href="02-fused-softmax.html#sphx-glr-getting-started-tutorials-02-fused-softmax-py"><span class="std std-ref">Fused Softmax</span></a> (<code class="docutils literal notranslate"><span class="pre">02-fused-softmax.py</span></code>)</p></td>
|
||||
<td><p>03:32.107</p></td>
|
||||
<td><p>03:30.914</p></td>
|
||||
<td><p>0.0 MB</p></td>
|
||||
</tr>
|
||||
<tr class="row-even"><td><p><a class="reference internal" href="01-vector-add.html#sphx-glr-getting-started-tutorials-01-vector-add-py"><span class="std std-ref">Vector Addition</span></a> (<code class="docutils literal notranslate"><span class="pre">01-vector-add.py</span></code>)</p></td>
|
||||
<td><p>01:50.498</p></td>
|
||||
<td><p>01:50.715</p></td>
|
||||
<td><p>0.0 MB</p></td>
|
||||
</tr>
|
||||
<tr class="row-odd"><td><p><a class="reference internal" href="04-low-memory-dropout.html#sphx-glr-getting-started-tutorials-04-low-memory-dropout-py"><span class="std std-ref">Low-Memory Dropout</span></a> (<code class="docutils literal notranslate"><span class="pre">04-low-memory-dropout.py</span></code>)</p></td>
|
||||
<td><p>00:00.282</p></td>
|
||||
<td><p>0.0 MB</p></td>
|
||||
</tr>
|
||||
<tr class="row-even"><td><p><a class="reference internal" href="07-libdevice-function.html#sphx-glr-getting-started-tutorials-07-libdevice-function-py"><span class="std std-ref">Libdevice function</span></a> (<code class="docutils literal notranslate"><span class="pre">07-libdevice-function.py</span></code>)</p></td>
|
||||
<td><p>00:00.250</p></td>
|
||||
<td><p>0.0 MB</p></td>
|
||||
</tr>
|
||||
<tr class="row-odd"><td><p><a class="reference internal" href="06-fused-attention.html#sphx-glr-getting-started-tutorials-06-fused-attention-py"><span class="std std-ref">Fused Attention</span></a> (<code class="docutils literal notranslate"><span class="pre">06-fused-attention.py</span></code>)</p></td>
|
||||
<td><p>00:00.073</p></td>
|
||||
<td><p>0.0 MB</p></td>
|
||||
</tr>
|
||||
<tr class="row-even"><td><p><a class="reference internal" href="04-low-memory-dropout.html#sphx-glr-getting-started-tutorials-04-low-memory-dropout-py"><span class="std std-ref">Low-Memory Dropout</span></a> (<code class="docutils literal notranslate"><span class="pre">04-low-memory-dropout.py</span></code>)</p></td>
|
||||
<td><p>00:00.012</p></td>
|
||||
<td><p>0.0 MB</p></td>
|
||||
</tr>
|
||||
<tr class="row-odd"><td><p><a class="reference internal" href="07-libdevice-function.html#sphx-glr-getting-started-tutorials-07-libdevice-function-py"><span class="std std-ref">Libdevice function</span></a> (<code class="docutils literal notranslate"><span class="pre">07-libdevice-function.py</span></code>)</p></td>
|
||||
<td><p>00:00.010</p></td>
|
||||
<td><p>00:00.074</p></td>
|
||||
<td><p>0.0 MB</p></td>
|
||||
</tr>
|
||||
</tbody>
|
||||
|
||||
File diff suppressed because one or more lines are too long
@@ -1,4 +1,4 @@
|
||||
# Sphinx build info version 1
|
||||
# This file hashes the configuration used when building these files. When it is not found, a full rebuild will be done.
|
||||
config: 67b632ed73e130ff238d935883579ed2
|
||||
config: b6a148ad6f1fecfeeccdf8b46659d01a
|
||||
tags: 645f666f9bcd5a90fca523b33c5a78b7
|
||||
|
||||
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user