• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

New SMP Core's

LINUX BIGADV!!!! YEAH!!!! I hope it comes out while I still have the DL580 G7 to try it out on! 32 native threads 64 with HT doing 2 GHZ.
 
New core raises specters. I'll let others be the ones to test it out.
 
I for one am looking forward to this. Don't have any boxes that will take advantage of linux -bigadv at present, but I might eventually.
 
LINUX BIGADV!!!! YEAH!!!! I hope it comes out while I still have the DL580 G7 to try it out on! 32 native threads 64 with HT doing 2 GHZ.

get on the beta team and put it to work...
 
New core raises specters. I'll let others be the ones to test it out.

i'm 100% with you there! Let the others who are always dying to be first to go ahead and try it out. Should the public beta test go rather well, then i'll jump on after a while.

From the sound what what Kasson is telling us, I guess i'll jump over when the A3 bigadv starts to dry up.
 
Oh its working right now just cranking SMP units and even with that it still kicks out at least 60k PPD.
 
The other big benefactor here is probably Tigerbiten because last I knew he was on all Linux also. I think a good percentage of his hardware is bigadv capable just their lack of bigadv on Linux was holding him back. I am a small player in this realm but if Tiger is still configured the way I heard last he will gain some ppd.
 
The other big benefactor here is probably Tigerbiten because last I knew he was on all Linux also. I think a good percentage of his hardware is bigadv capable just their lack of bigadv on Linux was holding him back. I am a small player in this realm but if Tiger is still configured the way I heard last he will gain some ppd.

gain 15k per bigadv unit turned in...
linux rocks... just sayin ... :)
 
Glad I had stopped folding yesterday - I finished my renders and noticed the new core release just before I fired them up. Working 2 new 6901s. :D
 
I have the 580 G7 for about two more weeks. I don't feel like joining the beta team just to massively increase my points though. As Patriot probably knows that G7 machine is seriously fast. SMP work units are mostly at or under a minute per frame.
 
I have the 580 G7 for about two more weeks. I don't feel like joining the beta team just to massively increase my points though. As Patriot probably knows that G7 machine is seriously fast. SMP work units are mostly at or under a minute per frame.


maybe you could work something out with patriot to run the system under his name. though he would probably have to ask the beta team permission to do it.
 
Mtnduey is the one who will benefit from Linux bigadv the most I believe. Tiger is all WinXP last I knew.
 
In super exciting new-core news, there is a new log entry mapping to 24 threads: ;)

original.jpg
 
well considering bigadv isn't even available to me right now it will be a nice boost.

lol.... yeah... I went from 150k to 220 by converting a few folders to windows bigadv... then I switched all machines back to linux as soon as I got into the beta team and am running bigadv...

given the lack of units I am guessing something hung... will check in the morning... but so long as I dont get the godforsaken 2684 units... my ppd is ~650k :eek::D
 
Ran a NUMA enabled test - still fucked. (12mins10sec vs 11mins15 sec)

Further frame times are utterly identical to 6900s. So no detectable performance improvements at all on the windows 32 bit version....

Anyone on Linux 64 vet?
 
Last edited:
All I run is Linux 64....

Using mostly rhel5u5 x64bit... new installs are rhel5u6 x64

I had to go read some guides to get windows to work right :eek:
Thankfully I am back in native territory...
 
Sorry I meant anyone tried the new client on linux 64?

Go update mid unit - you have machines to spare.;)
 
Further frame times are utterly identical to 6900s. So no detectable performance improvements at all on the windows 32 bit version....
No reason for me to convert right away if I stay with Windows then. Another possibility: I still have most of my Linux VMs intact and could relaunch them as soon as my current WUs complete. I don't even know if I should bother though, because of the performance overhead cost with a VM and it might just turn out to be similar in the end. It's a problem for those with mixed client systems in deciding the best course to take. To switch to Linux or not to switch. Perhaps this is fitting it's happening now since my exit time is approaching rather quickly...
 
10:27am new client released
10:59am first p6901 downloaded
11:24am first performance data posted
11:41am Unit captured and sent to [H] Benchmarking team
12:43pm NUMA tests complete.

[H] - your Folding@Home Fast Incident Response Centre. :p
 
No reason for me to convert right away if I stay with Windows then. Another possibility: I still have most of my Linux VMs intact and could relaunch them as soon as my current WUs complete. I don't even know if I should bother though, because of the performance overhead cost with a VM and it might just turn out to be similar in the end. It's a problem for those with mixed client systems in deciding the best course to take. To switch to Linux or not to switch. Perhaps this is fitting it's happening now since my exit time is approaching rather quickly...
There is no performance advantage to switching to Linux, nor will there probably ever be. This is just so that the people who run Linux natively for whatever reason or who don't have enough Windows licenses are still able to run -bigadv on any machine that has sufficient hardware to do so.
 
There is no advantage to switching to Linux, nor will there probably ever be. This is just so that the people who run Linux natively for whatever reason or who don't have enough Windows licenses are still able to run -bigadv on any machine that has sufficient hardware to do so.
There have been a few posts about superior performance on the same machines between the different OSes. These results must have been specific to beta PPD performance and not applicable to general release WUs then...?
 
Are you talking about posts on this forum? People have mentioned performance improvements going to Windows from Linux, but that's because Linux did not previously support -bigadv so they were going from regular SMP to -bigadv as well. The last time I used Linux, the A3 core performed better under Windows, and that was the case for A3 -bigadv as well when it was initially available for Linux.
 
Are you talking about posts on this forum?
Yup, and without naming names the posts were made by beta testers. Either the references made were only applicable to beta WUs or I totally misunderstood the context of the posts. In any case, if the performance difference is insignificant, no one who is currently running Windows -bigadv should even contemplate migrating to Linux unless there is another reason such as what you mentioned.
 
Yup, and without naming names the posts were made by beta testers. Either the references made were only applicable to beta WUs or I totally misunderstood the context of the posts. In any case, if the performance difference is insignificant, no one who is currently running Windows -bigadv should even contemplate migrating to Linux unless there is another reason such as what you mentioned.
I've seen the same posts
 
Yup, and without naming names the posts were made by beta testers. Either the references made were only applicable to beta WUs or I totally misunderstood the context of the posts. In any case, if the performance difference is insignificant, no one who is currently running Windows -bigadv should even contemplate migrating to Linux unless there is another reason such as what you mentioned.

non beta units fold faster on linux. because I blew away the partitions I don't have the previous frame times...
difference is not in seconds...

I will read up on the guidelines and see if those tpf on non beta units can be disclosed or not...
because the core is beta I don't know how that works...

The last time I used Linux, the A3 core performed better under Windows, and that was the case for A3 -bigadv as well when it was initially available for Linux.

When bigadv was first open on linux I heard talks of kernels that were slower to fold on... you may have been using one of the bad ones...
I should probably look those up and see if I can optimize my setups...


no one who is currently running Windows -bigadv should even contemplate migrating to Linux unless there is another reason such as what you mentioned.

network switch was causing issues... so we took the time to reboot some of the key servers... note... they were doing just fine...
539 days uptime...
 
When bigadv was first open on linux I heard talks of kernels that were slower to fold on... you may have been using one of the bad ones...
I should probably look those up and see if I can optimize my setups...
I never folded -bigadv under Linux. I folded SMP units under both Fedora 11 and Ubuntu 10.4, and both were slower than folding under Windows. Once the A3 core was available for Windows, I stopped using Linux altogether.
 
Anyone has problems downloading the A5 Core ? Im keep getting 404 errors.

A5Core.JPG
 
Is this new core / client works on WinXP 64bits yet? or it only works on Linux ?
 
Is this new core / client works on WinXP 64bits yet? or it only works on Linux ?

There is a Windows version that should work on XP 64 without issue. It is not just Linux.
 
I never folded -bigadv under Linux. I folded SMP units under both Fedora 11 and Ubuntu 10.4, and both were slower than folding under Windows. Once the A3 core was available for Windows, I stopped using Linux altogether.

deff bad kernel... I folded linux smp back in the days of c2q... 4k out of stock single p systems...

got ~1500k more than same clocked windows boxen...
 
deff bad kernel... I folded linux smp back in the days of c2q... 4k out of stock single p systems...

got ~1500k more than same clocked windows boxen...
That was probably before A3 units. I think it was pretty well known that Windows A3 was faster than Linux A3
Posted via Mobile Device
 
Back
Top