Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Thu, 13 Mar 2014 23:05:14 +0400
From: Solar Designer <solar@...nwall.com>
To: crypt-dev@...ts.openwall.com
Cc: Christian Buchner <Christian.Buchner@...il.com>
Subject: Re: using scrypt for user authentication

On Wed, Mar 12, 2014 at 11:13:27PM +0400, Solar Designer wrote:
> It's also important to note that only 2.5 GiB out of 6 GiB of TITAN's
> memory is in use.  (I could make the miner use more memory with other
> settings, but this is what appeared to provide optimal speed now, not
> needing to use more memory yet.)  Thus, I think the GPU/CPU speed ratio
> would remain about the same when YAC moves to higher Nfactor, for 8 MiB.
> 
> And the TITAN is likely to remain faster than its contemporary CPUs even
> at 16 MiB, and possibly at some larger sizes as well.  The current 4x
> performance gap is just too much to be fully removed by the move from
> 8 MiB to 16 MiB.  (Perhaps only a fraction of it would be removed.)

Today I've experimented with CudaMiner some further.  Specifically, I
used its offline benchmark feature along with its support for scrypt
proper (albeit limited to r=1).  (YAC uses a revision of scrypt with
ChaCha20 and Keccak in place of Salsa20 and SHA-256.)  I compared
CudaMiner's speeds on our Zotac GeForce GTX TITAN AMP! Edition against
those for scrypt compatibility mode in escrypt-0.3.2 (just to use an
optimized defensive implementation of scrypt) on our 2x E5-2670 with
8x DDR3-1600.

Here's a table with my results:

MiB | r | 1c/1t | 16c/32t | GPU ini | GPU stab | GPU MiB | GPU TMTO
-------------------------------------------------------------------
  1 | 8 |   700 |   15622 |         |          |         |
  1 | 1 |   655 |   14614 |  41.89k |   45.09k |    1895 | 2
  2 | 8 |   357 |    7665 |         |          |         |
  2 | 1 |   327 |    6282 |  15.23k |   14.80k |    1895 | 4
  4 | 8 |   179 |    3750 |         |          |         |
  4 | 1 |   164 |    2965 |   5.64k |    5.04k |    2511 | 6
  8 | 8 |    90 |    1886 |         |          |         |
  8 | 1 |    82 |    1443 |   2.18k |    1.63k |    3687 | 8
 16 | 8 |    45 |     945 |         |          |         |
 16 | 1 |    41 |     705 |   0.90k |    0.44k |    5367 | 11
 32 | 8 |    21 |     465 |         |          |         |
 32 | 1 |    17 |     347 |   0.28k |    0.14k |    6139 | 19
 64 | 8 |    10 |     210 |         |          |         |
 64 | 1 |     7 |     156 |   0.07k |    0.04k |    6139 | 38

The first column is scrypt's memory usage (per hash computed), the
second column is scrypt's "r" parameter (r=8 means 1 KiB blocks, r=1
means 128-byte blocks), the "1c/1t" column gives hashes/s speeds on
one CPU core running one thread (at 3.3 GHz due to turbo), and the
"16c/32t" column gives cumulative hashes/s speeds for all 16 cores
running 32 threads (HT is enabled, CPU clock rate is between 2.6 GHz and
3.0 GHz due to turbo).

The "GPU ini" column gives the hashes/s speeds that "cudaminer
--benchmark" reports right upon startup, whereas the "GPU stab" column
gives the speeds it reports after it's been running for a while.  I am
puzzled as to why these sometimes differ so much.  More on this below.

"GPU MiB" is the GPU card's memory usage while running the benchmark,
out of a maximum of 6 GiB installed on this card.  (Yes, the last two
benchmarks get very close to fully using the card's memory.  I chose the
TMTO factor that way, and frankly I didn't bother trying much higher
TMTO factor values that would result in lower memory usage.)  "GPU TMTO"
is the TMTO factor I used in these benchmarks.  (I tuned it manually,
except in the last two benchmarks where I merely set it to the lowest
value that fits.)

It can be seen that the GPU wins per-chip at up to 16 MiB, possibly
inclusive:

MiB | r | 1c/1t | 16c/32t | GPU ini | GPU stab | GPU MiB | GPU TMTO
-------------------------------------------------------------------
 16 | 8 |    45 |     945 |         |          |         |
 16 | 1 |    41 |     705 |   0.90k |    0.44k |    5367 | 11

The defensive speed is 705 or 945 hashes/s on _two_ server CPU chips
(so 352 to 472 per chip), whereas the attack speed is 440 to 900
hashes/s on _one_ gaming GPU.  It's the uncertainty between the two GPU
speeds reported that makes it unclear whether it still wins at 16 MiB.

At above 16 MiB per hash, we bump into the card's available memory at
what would otherwise be a likely more optimal TMTO factor, so we have to
start increasing the TMTO factor for that reason.

While I am not 100% confident these results are correct, it appears that
CudaMiner does perform some verification - e.g., when I set the TMTO
factor too low and end up exceeding the GPU card's memory, I get a flood
of nice "result does not validate on CPU" messages from CudaMiner.

As to the two GPU speeds, I think the higher speed reported initially
might actually be more correct in terms of what an attacker would be
able to achieve (and I don't know what causes much lower speeds to be
reported later).  I think so because I see a similar effect with YAC
mining benchmark:

$ ./cudaminer --algo=scrypt-jane:YAC -i 0 -b 8192 -L 6 -l t14x32 --benchmark
[...]
[2014-03-13 20:54:29] GPU #0: GeForce GTX TITAN, 6.82 khash/s
[2014-03-13 20:54:29] Total: 6.82 khash/s
[2014-03-13 20:54:35] GPU #0: GeForce GTX TITAN, 6.29 khash/s
[2014-03-13 20:54:35] Total: 6.29 khash/s
[2014-03-13 20:54:40] GPU #0: GeForce GTX TITAN, 6.21 khash/s
[2014-03-13 20:54:40] Total: 6.21 khash/s
[2014-03-13 20:54:45] GPU #0: GeForce GTX TITAN, 6.21 khash/s
[2014-03-13 20:54:45] Total: 6.21 khash/s
[2014-03-13 20:54:50] GPU #0: GeForce GTX TITAN, 6.20 khash/s
[2014-03-13 20:54:50] Total: 6.20 khash/s
[2014-03-13 20:54:56] GPU #0: GeForce GTX TITAN, 6.20 khash/s
[2014-03-13 20:54:56] Total: 6.20 khash/s
[2014-03-13 20:55:01] GPU #0: GeForce GTX TITAN, 6.20 khash/s
[2014-03-13 20:55:01] Total: 6.20 khash/s
[2014-03-13 20:55:06] GPU #0: GeForce GTX TITAN, 6.19 khash/s
[2014-03-13 20:55:06] Total: 6.19 khash/s

yet during actual mining, the speed stays at around 6.80 khash/s, and
the shares are being accepted by the pool just fine:

[2014-03-13 22:49:03] LONGPOLL detected new block
[2014-03-13 22:49:03] GPU #0: GeForce GTX TITAN, 6.73 khash/s
[2014-03-13 22:50:02] GPU #0: GeForce GTX TITAN, 6.83 khash/s
[2014-03-13 22:50:30] LONGPOLL detected new block
[2014-03-13 22:50:30] GPU #0: GeForce GTX TITAN, 6.87 khash/s
[2014-03-13 22:51:29] GPU #0: GeForce GTX TITAN, 6.83 khash/s
[2014-03-13 22:51:30] accepted: 57/57 (100.00%), 6.83 khash/s (yay!!!)

So I suspect that it's a benchmarking issue.

Below is some raw data I collected and used for the table above:

1 MiB (r = 8):

Benchmarking 1 thread ...
700 c/s real, 705 c/s virtual (1023 hashes in 1.46 seconds)
Benchmarking 32 threads ...
15622 c/s real, 490 c/s virtual (31713 hashes in 2.03 seconds)

1 MiB (r = 1):

Benchmarking 1 thread ...
655 c/s real, 660 c/s virtual (1023 hashes in 1.56 seconds)
Benchmarking 32 threads ...
14614 c/s real, 458 c/s virtual (15345 hashes in 1.05 seconds)

./cudaminer --algo=scrypt:8192 -i 0 -b 8192 -L 2 -l t14x32 --benchmark

[2014-03-13 20:39:22] GPU #0: GeForce GTX TITAN, 41.89 khash/s
[2014-03-13 20:39:22] Total: 41.89 khash/s
[2014-03-13 20:39:26] GPU #0: GeForce GTX TITAN, 45.37 khash/s
[2014-03-13 20:39:26] Total: 45.37 khash/s
[2014-03-13 20:39:31] GPU #0: GeForce GTX TITAN, 45.41 khash/s
[2014-03-13 20:39:31] Total: 45.41 khash/s
[2014-03-13 20:39:36] GPU #0: GeForce GTX TITAN, 45.41 khash/s
[2014-03-13 20:39:36] Total: 45.41 khash/s
[2014-03-13 20:39:42] GPU #0: GeForce GTX TITAN, 45.24 khash/s
[2014-03-13 20:39:42] Total: 45.24 khash/s
[2014-03-13 20:39:47] GPU #0: GeForce GTX TITAN, 45.09 khash/s
[2014-03-13 20:39:47] Total: 45.09 khash/s

+------------------------------------------------------+                       
| NVIDIA-SMI 5.319.60   Driver Version: 319.60         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TITAN   Off  | 0000:83:00.0     N/A |                  N/A |
| 40%   65C  N/A     N/A /  N/A |     1895MB /  6143MB |     N/A      Default |
+-------------------------------+----------------------+----------------------+

2 MiB (r = 8):

Benchmarking 1 thread ...
357 c/s real, 357 c/s virtual (511 hashes in 1.43 seconds)
Benchmarking 32 threads ...
7665 c/s real, 241 c/s virtual (7665 hashes in 1.00 seconds)

2 MiB (r = 1):

Benchmarking 1 thread ...
327 c/s real, 327 c/s virtual (511 hashes in 1.56 seconds)
Benchmarking 32 threads ...
6282 c/s real, 197 c/s virtual (7665 hashes in 1.22 seconds)

./cudaminer --algo=scrypt:16384 -i 0 -b 8192 -L 4 -l t14x32 --benchmark

[2014-03-13 20:42:25] GPU #0: GeForce GTX TITAN, 15.23 khash/s
[2014-03-13 20:42:25] Total: 15.23 khash/s
[2014-03-13 20:42:31] GPU #0: GeForce GTX TITAN, 15.09 khash/s
[2014-03-13 20:42:31] Total: 15.09 khash/s
[2014-03-13 20:42:36] GPU #0: GeForce GTX TITAN, 15.03 khash/s
[2014-03-13 20:42:36] Total: 15.03 khash/s
[2014-03-13 20:42:41] GPU #0: GeForce GTX TITAN, 14.97 khash/s
[2014-03-13 20:42:41] Total: 14.97 khash/s
[2014-03-13 20:42:46] GPU #0: GeForce GTX TITAN, 14.94 khash/s
[2014-03-13 20:42:46] Total: 14.94 khash/s
[2014-03-13 20:42:51] GPU #0: GeForce GTX TITAN, 14.92 khash/s
[2014-03-13 20:42:51] Total: 14.92 khash/s
[2014-03-13 20:42:56] GPU #0: GeForce GTX TITAN, 14.91 khash/s
[2014-03-13 20:42:56] Total: 14.91 khash/s
[2014-03-13 20:43:01] GPU #0: GeForce GTX TITAN, 14.90 khash/s
[2014-03-13 20:43:01] Total: 14.90 khash/s
[2014-03-13 20:43:06] GPU #0: GeForce GTX TITAN, 14.90 khash/s
[2014-03-13 20:43:06] Total: 14.90 khash/s
[2014-03-13 20:43:11] GPU #0: GeForce GTX TITAN, 14.90 khash/s
[2014-03-13 20:43:11] Total: 14.90 khash/s

[...]

[2014-03-13 20:43:57] Total: 14.84 khash/s
[2014-03-13 20:44:02] GPU #0: GeForce GTX TITAN, 14.81 khash/s
[2014-03-13 20:44:02] Total: 14.81 khash/s
[2014-03-13 20:44:07] GPU #0: GeForce GTX TITAN, 14.82 khash/s
[2014-03-13 20:44:07] Total: 14.82 khash/s
[2014-03-13 20:44:12] GPU #0: GeForce GTX TITAN, 14.81 khash/s
[2014-03-13 20:44:12] Total: 14.81 khash/s
[2014-03-13 20:44:17] GPU #0: GeForce GTX TITAN, 14.80 khash/s
[2014-03-13 20:44:17] Total: 14.80 khash/s
[2014-03-13 20:44:22] GPU #0: GeForce GTX TITAN, 14.80 khash/s
[2014-03-13 20:44:22] Total: 14.80 khash/s
[2014-03-13 20:44:27] GPU #0: GeForce GTX TITAN, 14.79 khash/s
[2014-03-13 20:44:27] Total: 14.79 khash/s

[...]

[2014-03-13 20:45:03] GPU #0: GeForce GTX TITAN, 14.79 khash/s
[2014-03-13 20:45:03] Total: 14.79 khash/s
[2014-03-13 20:45:08] GPU #0: GeForce GTX TITAN, 14.80 khash/s
[2014-03-13 20:45:08] Total: 14.80 khash/s

+------------------------------------------------------+                       
| NVIDIA-SMI 5.319.60   Driver Version: 319.60         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TITAN   Off  | 0000:83:00.0     N/A |                  N/A |
| 49%   76C  N/A     N/A /  N/A |     1895MB /  6143MB |     N/A      Default |
+-------------------------------+----------------------+----------------------+

4 MiB (r = 8):

Benchmarking 1 thread ...
179 c/s real, 179 c/s virtual (255 hashes in 1.42 seconds)
Benchmarking 32 threads ...
3750 c/s real, 118 c/s virtual (3825 hashes in 1.02 seconds)

4 MiB (r = 1):

Benchmarking 1 thread ...
164 c/s real, 164 c/s virtual (255 hashes in 1.55 seconds)
Benchmarking 32 threads ...
2965 c/s real, 93 c/s virtual (3825 hashes in 1.29 seconds)

./cudaminer --algo=scrypt:32768 -i 0 -b 8192 -L 6 -l t14x32 --benchmark

[2014-03-13 20:46:02] GPU #0: GeForce GTX TITAN, 5.64 khash/s
[2014-03-13 20:46:02] Total: 5.64 khash/s
[2014-03-13 20:46:07] GPU #0: GeForce GTX TITAN, 5.08 khash/s
[2014-03-13 20:46:07] Total: 5.08 khash/s
[2014-03-13 20:46:13] GPU #0: GeForce GTX TITAN, 5.07 khash/s
[2014-03-13 20:46:13] Total: 5.07 khash/s
[2014-03-13 20:46:19] GPU #0: GeForce GTX TITAN, 5.06 khash/s
[2014-03-13 20:46:19] Total: 5.06 khash/s
[2014-03-13 20:46:24] GPU #0: GeForce GTX TITAN, 5.06 khash/s
[2014-03-13 20:46:24] Total: 5.06 khash/s
[2014-03-13 20:46:30] GPU #0: GeForce GTX TITAN, 5.06 khash/s
[2014-03-13 20:46:30] Total: 5.06 khash/s
[2014-03-13 20:46:36] GPU #0: GeForce GTX TITAN, 5.06 khash/s
[2014-03-13 20:46:36] Total: 5.06 khash/s
[2014-03-13 20:46:41] GPU #0: GeForce GTX TITAN, 5.05 khash/s
[2014-03-13 20:46:41] Total: 5.05 khash/s
[2014-03-13 20:46:47] GPU #0: GeForce GTX TITAN, 5.05 khash/s
[2014-03-13 20:46:47] Total: 5.05 khash/s

[...]

[2014-03-13 20:47:04] GPU #0: GeForce GTX TITAN, 5.04 khash/s
[2014-03-13 20:47:04] Total: 5.04 khash/s
[2014-03-13 20:47:10] GPU #0: GeForce GTX TITAN, 5.04 khash/s
[2014-03-13 20:47:10] Total: 5.04 khash/s

+------------------------------------------------------+                       
| NVIDIA-SMI 5.319.60   Driver Version: 319.60         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TITAN   Off  | 0000:83:00.0     N/A |                  N/A |
| 49%   76C  N/A     N/A /  N/A |     2511MB /  6143MB |     N/A      Default |
+-------------------------------+----------------------+----------------------+

8 MiB (r = 8):

Benchmarking 1 thread ...
90 c/s real, 90 c/s virtual (127 hashes in 1.40 seconds)
Benchmarking 32 threads ...
1886 c/s real, 59 c/s virtual (1905 hashes in 1.01 seconds)

8 MiB (r = 1):

Benchmarking 1 thread ...
82 c/s real, 83 c/s virtual (127 hashes in 1.54 seconds)
Benchmarking 32 threads ...
1443 c/s real, 45 c/s virtual (1905 hashes in 1.32 seconds)

./cudaminer --algo=scrypt:65536 -i 0 -b 8192 -L 8 -l t14x32 --benchmark

[2014-03-13 20:47:35] GPU #0: GeForce GTX TITAN, 2.18 khash/s
[2014-03-13 20:47:35] Total: 2.18 khash/s
[2014-03-13 20:47:44] GPU #0: GeForce GTX TITAN, 1.75 khash/s
[2014-03-13 20:47:44] Total: 1.75 khash/s
[2014-03-13 20:47:50] GPU #0: GeForce GTX TITAN, 1.64 khash/s
[2014-03-13 20:47:50] Total: 1.64 khash/s
[2014-03-13 20:47:57] GPU #0: GeForce GTX TITAN, 1.63 khash/s
[2014-03-13 20:47:57] Total: 1.63 khash/s
[2014-03-13 20:48:03] GPU #0: GeForce GTX TITAN, 1.64 khash/s
[2014-03-13 20:48:03] Total: 1.64 khash/s
[2014-03-13 20:48:10] GPU #0: GeForce GTX TITAN, 1.64 khash/s
[2014-03-13 20:48:10] Total: 1.64 khash/s
[2014-03-13 20:48:16] GPU #0: GeForce GTX TITAN, 1.63 khash/s
[2014-03-13 20:48:16] Total: 1.63 khash/s
[2014-03-13 20:48:23] GPU #0: GeForce GTX TITAN, 1.64 khash/s
[2014-03-13 20:48:23] Total: 1.64 khash/s

[...]

[2014-03-13 20:48:36] GPU #0: GeForce GTX TITAN, 1.63 khash/s
[2014-03-13 20:48:36] Total: 1.63 khash/s
[2014-03-13 20:48:43] GPU #0: GeForce GTX TITAN, 1.64 khash/s
[2014-03-13 20:48:43] Total: 1.64 khash/s
[2014-03-13 20:48:49] GPU #0: GeForce GTX TITAN, 1.63 khash/s
[2014-03-13 20:48:49] Total: 1.63 khash/s
[2014-03-13 20:48:56] GPU #0: GeForce GTX TITAN, 1.63 khash/s
[2014-03-13 20:48:56] Total: 1.63 khash/s

+------------------------------------------------------+                       
| NVIDIA-SMI 5.319.60   Driver Version: 319.60         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TITAN   Off  | 0000:83:00.0     N/A |                  N/A |
| 45%   74C  N/A     N/A /  N/A |     3687MB /  6143MB |     N/A      Default |
+-------------------------------+----------------------+----------------------+

16 MiB (r = 8):

Benchmarking 1 thread ...
45 c/s real, 45 c/s virtual (63 hashes in 1.39 seconds)
Benchmarking 32 threads ...
945 c/s real, 29 c/s virtual (945 hashes in 1.00 seconds)

16 MiB (r = 1):

Benchmarking 1 thread ...
41 c/s real, 41 c/s virtual (63 hashes in 1.53 seconds)
Benchmarking 32 threads ...
705 c/s real, 22 c/s virtual (945 hashes in 1.34 seconds)

./cudaminer --algo=scrypt:131072 -i 0 -b 8192 -L 11 -l t14x32 --benchmark

[2014-03-13 20:51:02] GPU #0: GeForce GTX TITAN, 0.90 khash/s
[2014-03-13 20:51:02] Total: 0.90 khash/s
[2014-03-13 20:51:14] GPU #0: GeForce GTX TITAN, 0.60 khash/s
[2014-03-13 20:51:14] Total: 0.60 khash/s
[2014-03-13 20:51:22] GPU #0: GeForce GTX TITAN, 0.45 khash/s
[2014-03-13 20:51:22] Total: 0.45 khash/s
[2014-03-13 20:51:30] GPU #0: GeForce GTX TITAN, 0.45 khash/s
[2014-03-13 20:51:30] Total: 0.45 khash/s
[2014-03-13 20:51:38] GPU #0: GeForce GTX TITAN, 0.44 khash/s
[2014-03-13 20:51:38] Total: 0.44 khash/s
[2014-03-13 20:51:46] GPU #0: GeForce GTX TITAN, 0.44 khash/s
[2014-03-13 20:51:46] Total: 0.44 khash/s
[2014-03-13 20:51:54] GPU #0: GeForce GTX TITAN, 0.44 khash/s
[2014-03-13 20:51:54] Total: 0.44 khash/s
[2014-03-13 20:52:02] GPU #0: GeForce GTX TITAN, 0.44 khash/s
[2014-03-13 20:52:02] Total: 0.44 khash/s
[2014-03-13 20:52:10] GPU #0: GeForce GTX TITAN, 0.44 khash/s
[2014-03-13 20:52:10] Total: 0.44 khash/s

+------------------------------------------------------+
| NVIDIA-SMI 5.319.60   Driver Version: 319.60         |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TITAN   Off  | 0000:83:00.0     N/A |                  N/A |
| 44%   73C  N/A     N/A /  N/A |     5367MB /  6143MB |     N/A      Default |
+-------------------------------+----------------------+----------------------+

32 MiB (r = 8):

Benchmarking 1 thread ...
21 c/s real, 21 c/s virtual (31 hashes in 1.41 seconds)
Benchmarking 32 threads ...
465 c/s real, 14 c/s virtual (465 hashes in 1.00 seconds)

32 MiB (r = 1):

Benchmarking 1 thread ...
17 c/s real, 17 c/s virtual (31 hashes in 1.79 seconds)
Benchmarking 32 threads ...
347 c/s real, 11 c/s virtual (465 hashes in 1.34 seconds)

./cudaminer --algo=scrypt:262144 -i 0 -b 8192 -L 20 -l t14x32 --benchmark

[2014-03-13 20:57:24] GPU #0: GeForce GTX TITAN, 0.28 khash/s
[2014-03-13 20:57:24] Total: 0.28 khash/s
[2014-03-13 20:57:50] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 20:57:50] Total: 0.14 khash/s
[2014-03-13 20:58:16] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 20:58:16] Total: 0.14 khash/s
[2014-03-13 20:58:42] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 20:58:42] Total: 0.14 khash/s
[2014-03-13 20:59:08] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 20:59:08] Total: 0.14 khash/s
[2014-03-13 20:59:35] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 20:59:35] Total: 0.14 khash/s
[2014-03-13 21:00:01] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 21:00:01] Total: 0.14 khash/s
[2014-03-13 21:00:27] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 21:00:27] Total: 0.14 khash/s
[2014-03-13 21:00:53] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 21:00:53] Total: 0.14 khash/s
[2014-03-13 21:01:19] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 21:01:19] Total: 0.14 khash/s
[2014-03-13 21:01:45] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 21:01:45] Total: 0.14 khash/s

+------------------------------------------------------+
| NVIDIA-SMI 5.319.60   Driver Version: 319.60         |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TITAN   Off  | 0000:83:00.0     N/A |                  N/A |
| 52%   77C  N/A     N/A /  N/A |     5871MB /  6143MB |     N/A      Default |
+-------------------------------+----------------------+----------------------+

./cudaminer --algo=scrypt:262144 -i 0 -b 8192 -L 19 -l t14x32 --benchmark

[2014-03-13 21:23:55] GPU #0: GeForce GTX TITAN, 0.28 khash/s
[2014-03-13 21:23:55] Total: 0.28 khash/s
[2014-03-13 21:24:20] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 21:24:20] Total: 0.14 khash/s
[2014-03-13 21:24:45] GPU #0: GeForce GTX TITAN, 0.14 khash/s
[2014-03-13 21:24:45] Total: 0.14 khash/s

+------------------------------------------------------+                       
| NVIDIA-SMI 5.319.60   Driver Version: 319.60         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TITAN   Off  | 0000:83:00.0     N/A |                  N/A |
| 51%   77C  N/A     N/A /  N/A |     6139MB /  6143MB |     N/A      Default |
+-------------------------------+----------------------+----------------------+

64 MiB (r = 8):

Benchmarking 1 thread ...
10 c/s real, 10 c/s virtual (15 hashes in 1.39 seconds)
Benchmarking 32 threads ...
210 c/s real, 6 c/s virtual (225 hashes in 1.07 seconds)

64 MiB (r = 1):

Benchmarking 1 thread ...
7 c/s real, 7 c/s virtual (15 hashes in 1.89 seconds)
Benchmarking 32 threads ...
156 c/s real, 5 c/s virtual (225 hashes in 1.44 seconds)

./cudaminer --algo=scrypt:524288 -i 0 -b 8192 -L 40 -l t14x32 --benchmark

[2014-03-13 21:04:28] GPU #0: GeForce GTX TITAN, 0.07 khash/s
[2014-03-13 21:04:28] Total: 0.07 khash/s
[2014-03-13 21:06:07] GPU #0: GeForce GTX TITAN, 0.04 khash/s
[2014-03-13 21:06:07] Total: 0.04 khash/s
[2014-03-13 21:07:46] GPU #0: GeForce GTX TITAN, 0.04 khash/s
[2014-03-13 21:07:46] Total: 0.04 khash/s
[2014-03-13 21:09:25] GPU #0: GeForce GTX TITAN, 0.04 khash/s
[2014-03-13 21:09:25] Total: 0.04 khash/s
[2014-03-13 21:11:04] GPU #0: GeForce GTX TITAN, 0.04 khash/s
[2014-03-13 21:11:04] Total: 0.04 khash/s
[2014-03-13 21:12:43] GPU #0: GeForce GTX TITAN, 0.04 khash/s
[2014-03-13 21:12:43] Total: 0.04 khash/s
[2014-03-13 21:14:22] GPU #0: GeForce GTX TITAN, 0.04 khash/s
[2014-03-13 21:14:22] Total: 0.04 khash/s

+------------------------------------------------------+                       
| NVIDIA-SMI 5.319.60   Driver Version: 319.60         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TITAN   Off  | 0000:83:00.0     N/A |                  N/A |
| 52%   77C  N/A     N/A /  N/A |     5871MB /  6143MB |     N/A      Default |
+-------------------------------+----------------------+----------------------+

./cudaminer --algo=scrypt:524288 -i 0 -b 8192 -L 38 -l t14x32 --benchmark

[2014-03-13 21:18:55] GPU #0: GeForce GTX TITAN, 0.07 khash/s
[2014-03-13 21:18:55] Total: 0.07 khash/s
[2014-03-13 21:20:30] GPU #0: GeForce GTX TITAN, 0.04 khash/s
[2014-03-13 21:20:30] Total: 0.04 khash/s
[2014-03-13 21:22:04] GPU #0: GeForce GTX TITAN, 0.04 khash/s
[2014-03-13 21:22:04] Total: 0.04 khash/s

+------------------------------------------------------+                       
| NVIDIA-SMI 5.319.60   Driver Version: 319.60         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TITAN   Off  | 0000:83:00.0     N/A |                  N/A |
| 52%   77C  N/A     N/A /  N/A |     6139MB /  6143MB |     N/A      Default |
+-------------------------------+----------------------+----------------------+

Alexander

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ