Woohoo! I got my first S5R4 task. :-D

koschi
koschi
Joined: 17 Mar 05
Posts: 86
Credit: 1655454430
RAC: 232001

My C2D E7200 @ 3.4GHz

My C2D E7200 @ 3.4GHz finished it's first S5R4 Unit :)

CPU type 2 was detected, which should mean that SSE2 was used, right?

http://einsteinathome.org/task/103563847

It took 19798.25 seconds, which is roughly double the time the fastest S5R3 Unit used to take with the Linux SSE2 app.
62.10 credits have been claimed and should then result in 181.3credits which gives a daily output of 1579. During S5R3 it was around 4000 credits / day...

All together thats ok with some room for optimizations :-D
My Q6600 @ 3,2GHz has problems to reach this 1580 Credits each day when crunching other projects like Spinhenge for example, which is really bad paid, especially under Linux, as they are unable to compile an app which is as fast as the Windows one. What ever, that is a different problem...

So it works and it's not too bad in credits.

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5779100
RAC: 0

RE: Anybody got any

Message 83122 in response to message 83119

Quote:
Anybody got any webspace with spare download bandwidth, that could host such a package?


You can always use Microsoft Skydrive. If you trust them. ;-)

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 2139
Credit: 2752801967
RAC: 1429758

RE: RE: Anybody got any

Message 83123 in response to message 83122

Quote:
Quote:
Anybody got any webspace with spare download bandwidth, that could host such a package?

You can always use Microsoft Skydrive. If you trust them. ;-)


OK, try this download link:

Einstein S5R4 Windows composite.zip.

The contents were OK when they left me, hosting is completely untested....

Only click if you trust me, and you also trust them.

Edit - if it works for you, could somebody with Linux please put together the same thing using approved Linux tools (all executable files from Gary's list, plus the app_info.xml, all in one .GZ archive), and I'll upload it to the same place. Email to initial dot surname at btinternet dot com.

koschi
koschi
Joined: 17 Mar 05
Posts: 86
Credit: 1655454430
RAC: 232001

Linux apps should be hosted

Linux apps should be hosted on Linux servers, not over at Microsoft ;-)

http://koschmider.de/files/einstein_linux_package_SSE.tar.gz
http://koschmider.de/files/einstein_linux_package_SSE2.tar.gz

You'll find in there the latest S5R3 4.49 and S5R4 6.02 apps for Linux and the app_info.xml. The first package contains the normal SSE application for S5R3, the second the SSE2 app. The S5R4 applications are the same, as 6.02 recognizes SSE2 capable hosts itself.

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284710168
RAC: 110462

Well done chaps. Cheers,

Well done chaps.

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

archae86
archae86
Joined: 6 Dec 05
Posts: 3145
Credit: 7023784931
RAC: 1804666

Hummm... Sometime in the

Hummm...

Sometime in the last few hours some parameter that BOINCMgr uses to estimate probable completion time for work units in the host queue changed drastically on one of my hosts, but not (yet?) on the other three.

In truth this host has been taking between four and 5 CPU hours for an S5R3 result using the 4.46 Windows ap, and the completion estimates have been in that range, bobbing up and down a bit with the sequence numbers of recently finished work.

But, very recently, something changed which has pushed that estimate for host Stoll4 up to almost 31 CPU hours for remaining S5R3, work, but just 8.5 hours for S5R4 work.

The Task duration correction factor currently shows as about .21, close to recent, so that is not what changed.

I've heard of lots of hosts downloading excessive S5R4 work at first try, so possibly there has been an intentional change, but the contrast in how it affected already downloaded S5R3 and S5R4 work on this host seems unhelpful.

If it is a one-off (as hinted by the lack of such effect on my other three hosts) this is unimportant, but if this will happen in general there will be much EDFing and perhaps some unfriendly project cooperation for participants running other projects.

[Edited to add that, yes, Stoll4 is running four Einstein tasks as "high priority" right now. No harm to me--I allocate 96% to Einstein anyway, but others would be less happy.]

[further edit to add that at first the S5R4 completion estimates were ludicrously shorter than truth, and they still are on my three unchanged hosts. So possibly a deliberate measure is propagating from the project to put S5R4 estimates closer to truth has as a side effect pushing S5R3 estimates through the tropopause--a flaw whose importance will drop quickly, with the fringe benefit of hastening the completion of S5R3 work)]

archae86
archae86
Joined: 6 Dec 05
Posts: 3145
Credit: 7023784931
RAC: 1804666

Somehow, despite having a

Somehow, despite having a large queue of S5R3, my host Stoll4 has completed, uploaded and reported its first S5R4 result

This took 8.53 CPU hours, while an S5R3 unit on this host ranged from about 4 to 5.

This unit was reported after BOINCMgr had increased the forecasted completion times for work already in queue by a factor of about six, but the host web page still showed the task duration correction factor at .21.

Subsequent to reporting this single S5R3 task, and a single S5R4 task, the completion forecasts have come down back down some (from almost 31 hours to 28.25 hours), and the task duration correction factor now shows on the web page as 1.30. (up from .21 minutes earlier)

So apparently through some means the tdcf was greatly increased, but the value displayed on the web page lagged that used by BOINCMgr to display predictions.

It seems likely that for hosts with large amounts of remaining S5R3 work in queue, that the hugely better than forecast performance will push the tdcf back down enough to escape the more severe levels of EDF and so on within a dozen or two task S5R3 completions.

However this may undo the (presumably intended) result of abating severe overfetch of S5R4 work once the S5R3 estimate excess bleeds off enough to allow fetch to resume, as it will mean that the tdcf has been driven back much nearer an appropriate value for S5R3, thus far too low for S5R3 as originally deployed. the tdcf, as I recall, will adapt upward immediately if the first S5R4 completion is substantially above expectation, but for fast hosts with large queues, by then the S5R4 work in queue may be several times that desired.

I invite correction from those who understand how this stuff really works.

tullio
tullio
Joined: 22 Jan 05
Posts: 2118
Credit: 61407735
RAC: 0

Finished my first S5R4 run in

Finished my first S5R4 run in 16 hours on my Linux box. Claimed credit 65.83, granted credit 192.24. The CPU is a 1.8 GHz AMD Opteron `1210 running SuSE Linux 10.3, BOINC 5.10.45, and is recognized as Type 2. Good work.
Tullio

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284710168
RAC: 110462

RE: Finished my first S5R4

Message 83129 in response to message 83128

Quote:
Finished my first S5R4 run in 16 hours on my Linux box. Claimed credit 65.83, granted credit 192.24. The CPU is a 1.8 GHz AMD Opteron `1210 running SuSE Linux 10.3, BOINC 5.10.45, and is recognized as Type 2. Good work.
Tullio


So that's ~ 192/16 ~ 12 credits / hour. So that compares with your most recent S3 unit ( #41418361 ) as ~ 237/6.26 ~ 38 credits /hour?

Hmmmm ....

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

tullio
tullio
Joined: 22 Jan 05
Posts: 2118
Credit: 61407735
RAC: 0

RE: RE: Finished my first

Message 83130 in response to message 83129

Quote:
Quote:
Finished my first S5R4 run in 16 hours on my Linux box. Claimed credit 65.83, granted credit 192.24. The CPU is a 1.8 GHz AMD Opteron `1210 running SuSE Linux 10.3, BOINC 5.10.45, and is recognized as Type 2. Good work.
Tullio

So that's ~ 192/16 ~ 12 credits / hour. So that compares with your most recent S3 unit ( #41418361 ) as ~ 237/6.26 ~ 38 credits /hour?

Hmmmm ....

Cheers, Mike.


QMC@home is giving me 22 c/h, SETI@home 17.8 c/h, CPDN 16 c/h. You pays your money and you takes your choice. But I still love Einstein, and lovers are blind.
Tullio

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.