Comments on First S5R3 Work Units

th3
th3
Joined: 24 Aug 06
Posts: 208
Credit: 2208434
RAC: 0

So The score so far: 15%

So The score so far:

15% slower on Linux, core2duo
2% slower on Windows, Core2quad (Tjeerd's comp)

Other R3 results on Linux, this is the current Top2 on E@H, dual Xeon 5355,
http://einsteinathome.org/host/795801,
how many percent slow down is this!
R3:
45,030.65 - 220.28
43,890.32 - 220.28
46,109.52 - 220.28
R2:
102,579.78 - 650.90
103,871.38 - 650.90
100,483.97 - 650.98

Looking bad for the penguins this run, like they dont have enough problems already with the polar caps melting and everything!

Udo
Udo
Joined: 19 May 05
Posts: 203
Credit: 8945570
RAC: 0

My first S5R3 finished and

My first S5R3 finished and validated ok:

h1_0541.80_S5R2__27_S5R2c in 299,025.58 Sec. with 653.29 credits - 457,72 sec/credit
h1_0541.80_S5R2__182_S5R3a in 119,813.22 Sec. with 220.28 credits - 543,91 sec/credit

Udo

Boinc_Master_2
Boinc_Master_2
Joined: 3 Mar 06
Posts: 25
Credit: 248324
RAC: 0

Got my first 4 yesterday,

Got my first 4 yesterday, I'll let you know.

Annika
Annika
Joined: 8 Aug 06
Posts: 720
Credit: 494410
RAC: 0

th3: *lol* nice way to put

th3: *lol* nice way to put it. The little devil/daemon gets a whacking too, as I read in the "S5R3"-thread... maybe an angel being a bit overly motivated...
Anyway, I'm having another penguin crunch right now (might be a guinea pig, really) to see if I can confirm the slowdown on my box.

Beezlebub
Beezlebub
Joined: 29 May 05
Posts: 9
Credit: 223322
RAC: 0

First one finished in 11hr

First one finished in 11hr 14min :)

e6600 quad @ 2.5ghz
2418 floating point
5227 integer

e6750 dual @ 3.71ghz
3657 floating point
8105 integer

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3522
Credit: 686139002
RAC: 550619

RE: My first S5R3 finished

Message 73177 in response to message 73173

Quote:

My first S5R3 finished and validated ok:

h1_0541.80_S5R2__27_S5R2c in 299,025.58 Sec. with 653.29 credits - 457,72 sec/credit
h1_0541.80_S5R2__182_S5R3a in 119,813.22 Sec. with 220.28 credits - 543,91 sec/credit

The absolute numbers appear kind of slow, I guess this is with HyperThreading?

CU

Bikeman

Udo
Udo
Joined: 19 May 05
Posts: 203
Credit: 8945570
RAC: 0

RE: RE: My first S5R3

Message 73178 in response to message 73177

Quote:
Quote:

My first S5R3 finished and validated ok:

h1_0541.80_S5R2__27_S5R2c in 299,025.58 Sec. with 653.29 credits - 457,72 sec/credit
h1_0541.80_S5R2__182_S5R3a in 119,813.22 Sec. with 220.28 credits - 543,91 sec/credit

The absolute numbers appear kind of slow, I guess this is with HyperThreading?

CU

Bikeman

yes, it's with hyperthreading!

Udo

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3522
Credit: 686139002
RAC: 550619

RE: RE: RE: My first

Message 73179 in response to message 73178

Quote:
Quote:
Quote:

My first S5R3 finished and validated ok:

h1_0541.80_S5R2__27_S5R2c in 299,025.58 Sec. with 653.29 credits - 457,72 sec/credit
h1_0541.80_S5R2__182_S5R3a in 119,813.22 Sec. with 220.28 credits - 543,91 sec/credit

The absolute numbers appear kind of slow, I guess this is with HyperThreading?

CU

Bikeman

yes, it's with hyperthreading!

Interesting.

Question to Bernd: is it possible that the new app are kind of less hyperthreading-friendly, because they are heavier on the FPU and have less integer stuff to mix with the floating point stuff? E.g. more FStat and less pattern matching, as a consequence to of the fewer sky grid points? Could also help to explain the compiler specific differences in performance.

CU

H-B

Eeyore69
Eeyore69
Joined: 2 Aug 07
Posts: 22
Credit: 608969
RAC: 0

RE: RE: RE: RE: My

Message 73180 in response to message 73179

Quote:
Quote:
Quote:
Quote:

My first S5R3 finished and validated ok:

h1_0541.80_S5R2__27_S5R2c in 299,025.58 Sec. with 653.29 credits - 457,72 sec/credit
h1_0541.80_S5R2__182_S5R3a in 119,813.22 Sec. with 220.28 credits - 543,91 sec/credit

The absolute numbers appear kind of slow, I guess this is with HyperThreading?

CU

Bikeman

yes, it's with hyperthreading!

Interesting.

Question to Bernd: is it possible that the new app are kind of less hyperthreading-friendly, because they are heavier on the FPU and have less integer stuff to mix with the floating point stuff? E.g. more FStat and less pattern matching, as a consequence to of the fewer sky grid points? Could also help to explain the compiler specific differences in performance.

CU

H-B


Eeyore69
Eeyore69
Joined: 2 Aug 07
Posts: 22
Credit: 608969
RAC: 0

I recently received a

I recently received a boatload of S5R3 WUs on my Intel Q6600 (overclocked to 3.4G) and running XP Pro. I have run a few for a comparison.

The S5R3 WUs take about 7.5 hours for 219 credits.

The S5R2 WUs took about 23 hours for 638 credits.

There seems to be no great difference on my system.

Eeyore

Comment viewing options

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