PDA

View Full Version : RC5 key rates...?



Bender10
01-22-2009, 04:19 PM
Could we get some 'key rate' info from those that have run RC5 CUDA..?

Include :

GPU description
OS
which 'core' you are running
key rate

and/or a '-bench' output...

My info:
9800GX2
Linux 64
Core 9
585 M/keys/s

Thanks

liuqyn
01-22-2009, 05:00 PM
9800GTX+
windows 64
core 9
340 Mkeys/s

vaughan
01-23-2009, 04:32 AM
8800GT
Linux Ubuntu 8.10 (64 bit)
Core 9
278 Mkeys / sec

[NGS]Cpt00Kirk
01-23-2009, 07:09 AM
Zotac 8800GT AMP Edition
windows XP SP3
core 9
335 Mkeys/s

vaughan
01-23-2009, 09:19 AM
Cpt00Kirk;63136']Zotac 8800GT AMP Edition
windows XP SP3
core 9
335 Mkeys/s
Wow that's better than 20 percent higher. Nice going [NGS]Cpt00Kirk.

I'd better take a look at why that is and as Captain Jean-Luc Picard says to Number One: 'Make it so'. :)

AMDave
01-23-2009, 09:49 AM
9600GT
Linux Ubuntu 8.10 (64 bit)
Core #0 1-pipe (previous version)
170 Mkeys / sec (previous version)

(previous version - something has gone haywire - I cannot run the bench with the last client - getting fatal errors now - possibly a driver config issue - I mess with stuff a lot - I think the key rate was about 173 to 175 Mkeys/sec before it went phut! - working on a fix will post new bench if I can)

Bender10
01-23-2009, 10:16 AM
@AmDave...I know. These rates are just ballpark numbers. Due to the changes from the first client to the one that just expired. I just wanted to get some of this down for basic information.

The next client should provide different rates (faster we hope..).

AMDave
01-23-2009, 11:08 AM
[Jan 23 12:03:53 UTC] RC5-72: using core #0 (CUDA 1-pipe 128-thd).
[Jan 23 12:04:12 UTC] RC5-72: Benchmark for core #0 (CUDA 1-pipe 128-thd)
0.00:00:16.19 [172,894,645 keys/sec]
[Jan 23 12:04:12 UTC] RC5-72: using core #1 (CUDA 2-pipe 64-thd).
[Jan 23 12:04:30 UTC] RC5-72: Benchmark for core #1 (CUDA 2-pipe 64-thd)
0.00:00:16.17 [174,743,283 keys/sec]
[Jan 23 12:04:30 UTC] RC5-72: using core #2 (CUDA 1-pipe 64-thd).
[Jan 23 12:04:50 UTC] RC5-72: Benchmark for core #2 (CUDA 1-pipe 64-thd)
0.00:00:16.93 [173,218,809 keys/sec]
[Jan 23 12:04:50 UTC] RC5-72: using core #3 (CUDA 1-pipe 256-thd).
[Jan 23 12:05:09 UTC] RC5-72: Benchmark for core #3 (CUDA 1-pipe 256-thd)
0.00:00:16.85 [169,830,487 keys/sec]
[Jan 23 12:05:09 UTC] RC5-72: using core #4 (CUDA 2-pipe 128-thd).
[Jan 23 12:05:29 UTC] RC5-72: Benchmark for core #4 (CUDA 2-pipe 128-thd)
0.00:00:16.85 [143,337,669 keys/sec]
[Jan 23 12:05:29 UTC] RC5-72: using core #5 (CUDA 1-pipe 64-thd sleep 100us).
[Jan 23 12:05:48 UTC] RC5-72: Benchmark for core #5 (CUDA 1-pipe 64-thd sleep 100us)
0.00:00:16.98 [172,703,159 keys/sec]
[Jan 23 12:05:48 UTC] RC5-72: using core #6 (CUDA 1-pipe 64-thd sleep dynamic).
[Jan 23 12:06:07 UTC] RC5-72: Benchmark for core #6 (CUDA 1-pipe 64-thd sleep dynamic)
0.00:00:16.93 [173,240,612 keys/sec]
[Jan 23 12:06:07 UTC] RC5-72: using core #7 (CUDA 1-pipe 64-thd busy wait).
[Jan 23 12:06:27 UTC] RC5-72: Benchmark for core #7 (CUDA 1-pipe 64-thd busy wait)
0.00:00:16.93 [173,161,770 keys/sec]
[Jan 23 12:06:27 UTC] RC5-72: using core #9 (CUDA 4-pipe 64-thd).
[Jan 23 12:06:46 UTC] RC5-72: Benchmark for core #9 (CUDA 4-pipe 64-thd)
0.00:00:16.42 [174,584,661 keys/sec]
[Jan 23 12:06:46 UTC] RC5-72: using core #10 (CUDA 4-pipe 128-thd).
[Jan 23 12:07:04 UTC] RC5-72: Benchmark for core #10 (CUDA 4-pipe 128-thd)
0.00:00:16.49 [147,781,319 keys/sec]

Bender10
01-23-2009, 11:36 AM
The reason these are just 'ballpark' information is due to the different 'key rates' reported in the client.

There are actually 3 different 'key rates'.

The rate reported by the -bench test. (this should be close to your unit completion time). This is a good estimate of your GPU throughput and is used by the client to select which 'core' (the fastest we hope..) to use.

The 'key rate' completion time you see listed when each unit is completed.
and...

The 'key rate average' that reports your running 'average' completion rate. This is usually lower than your actual completion rate.

[NGS]Cpt00Kirk
01-23-2009, 11:48 AM
Wow that's better than 20 percent higher. Nice going [NGS]Cpt00Kirk.

I'd better take a look at why that is and as Captain Jean-Luc Picard says to Number One: 'Make it so'. :)

its a factory standard overclock card : http://www.tweaktown.com/reviews/1262/1/zotac_geforce_8800gt_amp_edition/index.html

Brucifer
01-23-2009, 11:08 PM
ubuntu-64 8.10
260GTX
409
core 9

Bender10
01-24-2009, 01:37 PM
Thanks everyone for the input. This is all good info. I am going to try and get an admin to post this some how. Maybe as a link on our stats page...?

Brucifer
01-24-2009, 05:17 PM
Just be sure and post too that it all changes each time they put out a new beta client. :)

jamers
01-25-2009, 08:28 PM
9800 GTX+
core 9 Cuda 4 pipe 64 thd
393
Win XP_64

liuqyn
01-26-2009, 08:13 AM
new beta3 client
9800 GTX+ (stock speed)
core 6
390 Mkeys/s
win 64

AMDave
01-26-2009, 09:22 AM
new beta3 client
9600 GT (stock speed)
core 3
174 Mkeys/s
Ubuntu 8.10 AMD64

Brucifer
01-26-2009, 04:00 PM
Beta3
ubuntu-64 8.10
BFG 260GTX
454,212,648
core 7