Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date: Tue, 27 Mar 2012 23:26:58 +0800
From: myrice <qqlddg@...il.com>
To: john-dev@...ts.openwall.com
Subject: rawsha256.cu patch(using shared memory)

Hi,
Lukas, Solar, all

I used shared memory in rawsha256.cu(Just as Lukas comments as to-do)
There are still space for improvement. I think sha256 access patterns have
bank conflict.
Overall speedup by ~6% in sha256 and 8% in sha224
The file is in attachment and my space:
http://myrice.me/gsoc/rawsha256.cu.diff
http://myrice.me/gsoc/rawsha256.cu
Here are the test result:
Sha224:

=====Before===============
test: 5
format: raw-sha224-cuda
times: 10
Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1875K c/s real, 1920K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1881K c/s real, 1933K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1918K c/s real, 1944K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1860K c/s real, 1911K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1904K c/s real, 1938K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1872K c/s real, 1915K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1911K c/s real, 1949K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1894K c/s real, 1940K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1898K c/s real, 1929K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1840K c/s real, 1933K c/s virtual

Average: 1885.3 c/s real, 1931.2 c/s virtual

=======After=============

test: 5
format: raw-sha224-cuda
times: 10
Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 1976K c/s real, 2017K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 2024K c/s real, 2062K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 2098K c/s real, 2122K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 2069K c/s real, 2082K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 2067K c/s real, 2082K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 2087K c/s real, 2106K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 2057K c/s real, 2069K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 2046K c/s real, 2080K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 2077K c/s real, 2098K c/s virtual

Benchmarking: raw-sha224-cuda [SHA224]... DONE
Raw: 2069K c/s real, 2095K c/s virtual

Average: 2059.7 c/s real, 2081.3 c/s virtual

Speedup:~8%

Sha256:
=====Before===============

test: 5
format: raw-sha256-cuda
times: 10
Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 1931K c/s real, 1963K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 1931K c/s real, 1967K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 1902K c/s real, 1953K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 1920K c/s real, 1963K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 1956K c/s real, 1979K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 1924K c/s real, 1958K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 1915K c/s real, 1963K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 1922K c/s real, 1949K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 1953K c/s real, 1963K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 1979K c/s real, 1998K c/s virtual
Average: 1933.3 c/s real, 1965.6 c/s virtual

============After=================
test: 5
format: raw-sha256-cuda
times: 10
Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 2027K c/s real, 2067K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 2052K c/s real, 2074K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 2069K c/s real, 2100K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 2052K c/s real, 2100K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 2022K c/s real, 2067K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 2042K c/s real, 2064K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 2062K c/s real, 2098K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 2029K c/s real, 2042K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 2069K c/s real, 2103K c/s virtual

Benchmarking: raw-sha256-cuda [SHA256]... DONE
Raw: 2062K c/s real, 2085K c/s virtual
Average: 2048.6 c/s real, 2080.0 c/s virtual

Speedup: ~6%

Content of type "text/html" skipped

View attachment "rawsha256.cu.diff" of type "text/x-patch" (2363 bytes)

Download attachment "rawsha256.cu" of type "application/octet-stream" (3968 bytes)

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.