Ubuntu and no SSE

M

Mayhem33

[H]ard|Gawd
Joined
Oct 12, 2004
Messages
1,204
I am running Ubuntu Edgy 6.10 on my laptop and it is a Celeron 1.4 that has SSE 2 capability and I am only running with 3dnow support.

# Linux Console Edition #######################################################
###############################################################################

Folding@Home Client Version 5.04beta

http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: /root/foldingathome/CPU1
Executable: /root/foldingathome/CPU1/FAH504-Linux.exe
Arguments: -forceasm -verbosity 9

Warning:
By using the -forceasm flag, you are overriding
safeguards in the program. If you did not intend to
do this, please restart the program without -forceasm.
If work units are not completing fully (and particularly
if your machine is overclocked), then please discontinue
use of the flag.

[07:34:34] - Ask before connecting: No
[07:34:34] - User name: Mayhem33 (Team 33)
[07:34:34] - User ID: xxxxxxxxxxxxxxx
[07:34:34] - Machine ID: 1
[07:34:34]
[07:34:34] Loaded queue successfully.
[07:34:34] + Benchmarking ...
[07:34:52] The benchmark result is 1382
[07:34:52]
[07:34:52] + Processing work unit
[07:34:52] Core required: FahCore_a0.exe
[07:34:52] Core found.
[07:34:52] - Autosending finished units...
[07:34:52] Trying to send all finished work units
[07:34:52] + No unsent completed units remaining.
[07:34:52] - Autosend completed
[07:34:54] Working on Unit 02 [December 19 07:34:54]
[07:34:54] + Working ...
[07:34:54] - Calling './FahCore_a0.exe -dir work/ -suffix 02 -checkpoint 30 -forceasm -verbose -lifeline 4226 -version 504'

[07:34:55]
[07:34:55] *------------------------------*
[07:34:55] Folding@Home Gromacs 3.3 Core
[07:34:55] Version 1.91 (September 13, 2006)
[07:34:55]
[07:34:55] Preparing to commence simulation
[07:34:55] - Ensuring status. Please wait.
[07:35:12] - Assembly optimizations manually forced on.
[07:35:12] - Not checking prior termination.
[07:35:14] - Expanded 3630172 -> 16858811 (decompressed 464.4 percent)
[07:35:15]
[07:35:15] Project: 1487 (Run 0, Clone 546, Gen 4)
[07:35:15]
[07:35:15] Assembly optimizations on if available.
[07:35:15] Entering M.D.
[07:35:39] (Starting from checkpoint)
[07:35:39] Protein: p1487_DPPC_DOPC_CHOL
[07:35:39] Writing local files
[07:35:39] Completed 405871 out of 1000000 steps (40%)
[07:35:40] Extra 3DNow boost OK.
[07:35:41] Warning: long 1-4 interactions
[08:05:40] Timered checkpoint triggered.
[08:35:41] Timered checkpoint triggered.
[08:40:16] Writing local files
[08:40:16] Completed 410000 out of 1000000 steps (41%)
[09:10:17] Timered checkpoint triggered.
[09:40:18] Timered checkpoint triggered.
[10:10:19] Timered checkpoint triggered.
[10:40:20] Timered checkpoint triggered.
[11:10:20] Timered checkpoint triggered.
[11:16:02] Writing local files
[11:16:02] Completed 420000 out of 1000000 steps (42%)

I can not get the SSE enabled and do not understand what I am missing.
Help please!
 
It might just be that that Core doesn't use SSE. Some do, some don't.

Or I have no clue what I'm talking about....

[03:11:19] Core required: FahCore_a0.exe
[03:11:19] Core found.
[03:11:19] Working on Unit 09 [December 19 03:11:19]
[03:11:19] + Working ...
[03:11:20]
[03:11:20] *------------------------------*
[03:11:20] Folding@Home Gromacs 3.3 Core
[03:11:20] Version 1.71 (February 3, 2006)
[03:11:20]
[03:11:20] Preparing to commence simulation
[03:11:20] - Assembly optimizations manually forced on.
[03:11:20] - Not checking prior termination.
[03:11:23] - Expanded 3225247 -> 17508764 (decompressed 542.8 percent)
[03:11:24] - Starting from initial work packet
[03:11:24]
[03:11:24] Project: 1496 (Run 21, Clone 56, Gen 2)
[03:11:24]
[03:11:24] Assembly optimizations on if available.
[03:11:24] Entering M.D.
[03:11:33] Protein: p1496_Protein in POPC
[03:11:33] Writing local files
[03:11:37] Extra SSE boost OK.
 
My C2D is also working on a 1487 using ubuntu 6.1, and I don't show the SSE boost either:

[21:33:34] *------------------------------*
[21:33:34] Folding@Home Gromacs 3.3 Core
[21:33:34] Version 1.91 (September 13, 2006)
[21:33:34]
[21:33:34] Preparing to commence simulation
[21:33:34] - Ensuring status. Please wait.
[21:33:52] - Assembly optimizations manually forced on.
[21:33:52] - Not checking prior termination.
[21:33:53] - Expanded 3620998 -> 16858811 (decompressed 465.5 percent)
[21:33:53]
[21:33:53] Project: 1487 (Run 0, Clone 712, Gen 3)
[21:33:53]
[21:33:54] Assembly optimizations on if available.
[21:33:54] Entering M.D.
[21:34:16] (Starting from checkpoint)
[21:34:16] Protein: p1487_DPPC_DOPC_CHOL
[21:34:16] Writing local files
[21:34:16] Completed 817918 out of 1000000 steps (81%)
[21:34:17] Extra 3DNow boost OK.

Oddly enough, neither is a 1496 running on the other core:

[12:43:15] *------------------------------*
[12:43:15] Folding@Home Gromacs 3.3 Core
[12:43:15] Version 1.91 (September 13, 2006)
[12:43:15]
[12:43:15] Preparing to commence simulation
[12:43:15] - Assembly optimizations manually forced on.
[12:43:15] - Not checking prior termination.
[12:43:16] - Expanded 3209194 -> 17524417 (decompressed 546.0 percent)
[12:43:16] - Starting from initial work packet
[12:43:16]
[12:43:16] Project: 1496 (Run 2, Clone 4, Gen 10)
[12:43:16]
[12:43:16] Assembly optimizations on if available.
[12:43:16] Entering M.D.
[12:43:23] Protein: p1496_Protein in POPC
[12:43:23] Writing local files
[12:43:24] Extra 3DNow boost OK.
[12:43:25] Writing local files
[12:43:25] Completed 0 out of 500000 steps (0%)

 
Yes, I do.

This machine has been running a steady stream of 1487 and 1496 proteins, and I see no mention of a sse boost in any of the logs.
 
GameALot said:
Yes, I do.

This machine has been running a steady stream of 1487 and 1496 proteins, and I see no mention of a sse boost in any of the logs.
Click to expand...

Oldbenwa is running the same protein as you and his has the SSE going. This is what is making me crazy.

 
i think you guys are getting the 3dnow boost as opposed the SSE boost, which if im correct is close enough to the same thing. Cuz isnt 3dnow AMD and SSE intel? and heck, you are getting a boost from atleast something :)

i only say this cuz i know on a Mac, you get a Altivec boost. So i think it is whatever your processor's optimize code base is.


i may be wrong, but i think im right?
 
I wonder if it is a difference between the Linux and Windows clients.

Paging Dr. Oldbenwa. What client are you running?
 
GameALot said:
I wonder if it is a difference between the Linux and Windows clients.

Paging Dr. Oldbenwa. What client are you running?
Click to expand...
Paging Dr. Mage too since im sure he knows all this stuff and what the hell we are talking about
 
GameALot said:
I wonder if it is a difference between the Linux and Windows clients.

Paging Dr. Oldbenwa. What client are you running?
Click to expand...

Core required: FahCore_a0.exe
This is only used by Linux. He is running nix.

 
Boost you say?

vtec.png




I was under the impression that SSE and 3dNow didn't happen at the same time. I could be wrong.
 
Mayhem33 said:
I am running Ubuntu Edgy 6.10 on my laptop and it is a Celeron 1.4 that has SSE 2 capability and I am only running with 3dnow support.

[07:35:40] Extra 3DNow boost OK.

I can not get the SSE enabled and do not understand what I am missing.
Help please!
Click to expand...

Don't worry, your not missing anything, this is just a reporting glitch in the core.

FYI, Intel CPUs don't have, and hence can't use 3DNow+ optimizations.

If you want to prove to yourself that optimizations are being used to process the WU, just disable optimizations in the client configuration and you'll see your frame times are much longer.

ChelseaOilman

 
roftranspo said:
i think you guys are getting the 3dnow boost as opposed the SSE boost, which if im correct is close enough to the same thing. Cuz isnt 3dnow AMD and SSE intel? and heck, you are getting a boost from atleast something :)

i only say this cuz i know on a Mac, you get a Altivec boost. So i think it is whatever your processor's optimize code base is.


i may be wrong, but i think im right?
Click to expand...

Yes, you only get to run one or the other as far as 3DNow and SSE. The SSE produces more work than using the 3DNow. The processor is capable of SSE2 and i believe that SSE optimizations are supposed to be available for this protein. This is reason I am wondering what is going on.

 
ChelseaOilman said:
Don't worry, your not missing anything, this is just a reporting glitch in the core.

FYI, Intel CPUs don't have, and hence can't use 3DNow+ optimizations.

If you want to prove to yourself that optimizations are being used to process the WU, just disable optimizations in the client configuration and you'll see your frame times are much longer.

ChelseaOilman

Click to expand...

This is a bug and the frames really do take that long to complete. OK Thank you for explaining that to me/us! I will rest better knowing this.

 
AtomicMoose said:
Boost you say?

vtec.png




I was under the impression that SSE and 3dNow didn't happen at the same time. I could be wrong.
Click to expand...
man....why do you have to use that pic??? That was before i got the rims on that bad boy :D
 
You must log in or register to reply here.
Back
Top