Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 6 Jul 2015 20:46:12 +0800
From: Lei Zhang <zhanglei.april@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: tuning OMP_SCALE on MIC (was: Lei's weekly report #7)


> On Jun 24, 2015, at 9:53 AM, Solar Designer <solar@...nwall.com> wrote:
> 
>> dynamic_1023    [577188, 673663, 650376, 672000, 551764]
>> dahua           [21644, 42613, 75956, 136439, 212831]
>> Panama          [1284096, 1146880, 1819648, 2035712, 1872896]
>> skein-256       [1482752, 1631232, 2254848, 2530304, 2923520]
>> skein-512       [1440768, 1711104, 1936384, 2379776, 3183616]
>> HAVAL-256-3     [1345536, 1452032, 1657856, 2137088, 2437120]
>> Tiger           [1288192, 1433600, 1645568, 1949696, 2394112]
>> mdc2            [16717, 26401, 45405, 78721, 133338]
>> Raw-Keccak-256  [1235968, 1449984, 1841152, 2293760, 3083264]
>> HAVAL-128-4     [1486848, 1712128, 2068480, 2915328, 3301376]
>> ripemd-128      [1531904, 1748992, 2014208, 2372608, 2898944]
>> whirlpool       [988752, 1098752, 1144832, 1500160, 1599488]
>> ripemd-160      [1400832, 808941, 1857536, 2558976, 3176448]
>> Snefru-128      [934574, 951920, 1077248, 1479680, 1494016]
>> Raw-Keccak      [1060864, 1133568, 1645568, 1911808, 1969152]
>> has-160         [987089, 1037312, 1164288, 1376256, 1532928]
>> Snefru-256      [918178, 1142784, 1081344, 1432576, 1611776]
>> MD2             [547485, 612705, 735058, 872554, 933647]
>> VNC             [3179520, 4987904, 6769664, 8011776, 8549376]
>> MongoDB         [3656704, 5729280, 7598080, 9051136, 10071040]
>> OpenVMS         [3147776, 4302848, 5317632, 5926912, 6306816]
>> Raw-Blake2      [1141760, 1236992, 1360896, 1832960, 2039808]
> 
> Most of these are fast crap (even if they appear slow at this test).
> 
> Please sanity-check against speeds you obtain on CPU.  And no, I am not
> asking you to post more data in here - I am merely suggesting what
> checks to perform when you work on this.

To get a feeling of what performance to anticipate from MIC, I compared the single-threaded performance of every format between MIC and CPU (my laptop). Listed below are the "relatively" fastest formats and slowest formats on MIC:

Panama          0.025
Oracle12C       0.029
scrypt          0.036
django-scrypt   0.042
Clipperz,       0.051
(...)
mscash2,        0.318
sha1crypt,      0.328
Citrix_NS10,    0.332
xsha,           0.336
Office,         0.354

The numbers on the second column denote <performance on MIC>/<performance on CPU>. At its best, MIC runs about 1/3 as fast as CPU; at its worst, it's only 1/40 as fast. Formats on top of the list are obviously poorly optimized for MIC.


>> tc_whirlpool    [545, 685, 1177, 680, 680]
>> vtp             [464, 922, 1794, 3654, 7112]
>> keyring         [26713, 34285, 42666, 46900, 47627]
> 
> Out of them, tc_whirlpool and vtp show surprisingly low speeds, whereas
> keyring isn't that bad.  (Comparing to speeds seen on CPU.)

For those three formats, the ratios are 0.056, 0.148 and 0.285, respectively. Compared to the highest ratio of 0.354, vtp and keyring are not so bad as they seemed. BTW, MIC's benchmark result in this message is obtained with tuned OMP_SCALEs (not yet in my previous benchmarking).

OTOH, single-thread performance may not reflect scalability, so the relatively fast formats here are not necessarily fast when OMP_NUM_THREADS is increased to 240. But still, I think it's relevant to consider single-threaded performance of MIC. Perhaps I should first investigate those relatively slowest formats first, and then move on down the list.

In case it's useful, I attached the full list.


Lei


View attachment "perf-cmp.txt" of type "text/plain" (5020 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.