• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

SOB: New Windows client -- Fix for SSE2 problems

Just in time for the MTR 🙂

Unfortunately, it's not so much of a 'fix' as a 'sweep all the problems under the doormat' - as far as I know, SSE2 is disabled for high exponents on P4s, so there'll be a ~10% performance hit. 🙁
 
Originally posted by: Allio
Just in time for the MTR 🙂

Unfortunately, it's not so much of a 'fix' as a 'sweep all the problems under the doormat' - as far as I know, SSE2 is disabled for high exponents on P4s, so there'll be a ~10% performance hit. 🙁

Yeha I heard about that, sucks 🙁

Still, at least it will keep things running 🙂

Better get your fleet sorted Jeremy 😛
 
Originally posted by: Allio
Just in time for the MTR 🙂

Unfortunately, it's not so much of a 'fix' as a 'sweep all the problems under the doormat' - as far as I know, SSE2 is disabled for high exponents on P4s, so there'll be a ~10% performance hit. 🙁
Much more than a 10% hit on mine unless I did something wrong. Couple of points:

1.) The client window still shows 1.1.0 at the top

2.) When you open the About window it does show 1.1.1

3.) My main system when running 2 instances of the 1.1.0 client using -o2 would show at least 480xxx cEMs/sec on each client. With the 1.1.1 client having turned off SSE2 and using the -o2 I am only getting about 298xxx cEMs/sec. That's closer to a ~40% performance hit.

4.) My other P4 rig took about the same performance hit as my main rig.

5.) I used the following sobsvc switches:
sobsvc -i
sobsvc -o2
sobsvc -s
sobsvc -m
sobsvc -k

Anybody getting this type of performance hit besides me?

Do I need to use a different optimization?

 
jhites,

Not had a chance to check out my P4 system here at home, but I installed the latest client on my AXP rig, and it did still show 1.1.0 at the top of the client. Didn't check the About window.

I can't remember how fast my P4 system was going on other tests, but shall see how it does without SSE2, and report back. I know from past experiences, that P4s need SSE2 to be fast, otherwise they're not too brilliant. The P4 architecture was designed with a huge emphasis on SSE2 (hence why it was good when SoB came out with SSE2 enhancements 😉)


Garry
 
Originally posted by: Confused
jhites,

Not had a chance to check out my P4 system here at home, but I installed the latest client on my AXP rig, and it did still show 1.1.0 at the top of the client. Didn't check the About window.

I can't remember how fast my P4 system was going on other tests, but shall see how it does without SSE2, and report back. I know from past experiences, that P4s need SSE2 to be fast, otherwise they're not too brilliant. The P4 architecture was designed with a huge emphasis on SSE2 (hence why it was good when SoB came out with SSE2 enhancements 😉)


Garry


Hopefully this is just a quick fix to get all the P4s running and louie will release a client that fixes the SSE2 bug in a week or two. At least we can run under our own names now, not just the secret account...
 
Originally posted by: MemnochtheDevil
Hopefully this is just a quick fix to get all the P4s running and louie will release a client that fixes the SSE2 bug in a week or two. At least we can run under our own names now, not just the secret account...
Yes, according to the SoB Forum that is exactly what the quick fix is meant to do - Get them up and running.
Quote from SoB Forum
This is a short term fix to get all the P4s back up and working again. This gives Louie a bit of breathing space to get the SSE2 fix sorted.

 
I've now got all 13 clients that were stuck un-stuck with the SSE2-crippled 1.1.1

I'm gonna be getting some more production, but with 13 1.7Ghz+ machines running without SSE2, my rate's gonna be a hell of a lot lower until it's fixed 🙁 *cries* But, it's still going to be better for now than 13 clients not running.


Garry
 
I'm having the same problems with the new client as I was with the old one on my Pentium-M. Crashes as soon as it starts to process a new test with n value of 498K+.🙁
 
Back
Top