climateprediction.net (CPDN) home page
Thread 'Core 2 Duo E6600 speed'

Thread 'Core 2 Duo E6600 speed'

Message boards : Number crunching : Core 2 Duo E6600 speed
Message board moderation

To post messages, you must log in.

AuthorMessage
Profileold_user81594

Send message
Joined: 11 Jun 05
Posts: 67
Credit: 1,222,916
RAC: 0
Message 25428 - Posted: 2 Dec 2006, 16:08:15 UTC

Hi,

I\'ve just upgraded again from an Intel 950D (which was pretty impressive) to a Core 2 Duo E6600 (2.4GHz/4Mb L2 cache)

With CPDN and my old 950D, I was running at 1.9secs/TS. I\'ve just checked my current CPDN model over the last 3.5 hours, and it works out at only 1.68s/TS with my Core 2 Duo chip.
I was expecting a bit better than that. Could I just be seeing a \"slow\" part of the model?? Any others with an E6600 - what is you seconds/Timestep stat???

I thought I would be nearer to 1.3s/TS with this CPU.

Neil.
ID: 25428 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 25429 - Posted: 2 Dec 2006, 16:37:16 UTC

This is being discussed on the php boards here, and the quad core here.

There is also a thread here about a slow E6400.

ID: 25429 · Report as offensive     Reply Quote
Chidge

Send message
Joined: 7 Jun 06
Posts: 5
Credit: 3,228,997
RAC: 0
Message 25446 - Posted: 3 Dec 2006, 16:22:59 UTC

My e6600 gets 1.66 - 1.64 sec/TS on 1 core

link
ID: 25446 · Report as offensive     Reply Quote
Desti

Send message
Joined: 6 Aug 04
Posts: 124
Credit: 9,195,838
RAC: 0
Message 25447 - Posted: 3 Dec 2006, 18:30:45 UTC

I have

Average (sec/TS) 1.6484

with a E6400 @ 2662/333 MHz with DDR2 667 memory on a i945G graphic chipset.

CPDN is running on core 0 while core 1 is crunching Einstein@home.
Linux Users Everywhere @ BOINC
ID: 25447 · Report as offensive     Reply Quote
Profileold_user81594

Send message
Joined: 11 Jun 05
Posts: 67
Credit: 1,222,916
RAC: 0
Message 25468 - Posted: 5 Dec 2006, 18:30:24 UTC - in response to Message 25446.  

My e6600 gets 1.66 - 1.64 sec/TS on 1 core

link


Well, I\'m getting the same 1.62s/TS with both cores running CPDN. If I suspend my projects so I am just running one instance of CPDN on \"one\" core, I get 1.39s/TS.
My PC is E6600 on Asus P5WDH Deluxe with 2Gb OCZ DDR2 PC6400 (5-5-5-15) RAM.

Neil.
ID: 25468 · Report as offensive     Reply Quote
Helmer Bryd

Send message
Joined: 16 Aug 04
Posts: 156
Credit: 9,035,872
RAC: 2,928
Message 25530 - Posted: 9 Dec 2006, 13:02:37 UTC

These 1.10-1.30 speeds people report are overclocked machines, and the beauty about this chip is that they are rather easy to overclock.

Mine E6600 @3.2 GHz is doing 1.32-1.35 with two models and the Linux client (which is a couple of % slower than Windows)
With one model it\'s around 1.25.
With one model + a GIMPS M32 LL test it\'s around 1.45, hmm...

ID: 25530 · Report as offensive     Reply Quote
Helmer Bryd

Send message
Joined: 16 Aug 04
Posts: 156
Credit: 9,035,872
RAC: 2,928
Message 25533 - Posted: 9 Dec 2006, 16:38:27 UTC
Last modified: 9 Dec 2006, 16:40:42 UTC

Might to add:
One model + 1 Einstein task shows 1.26 s/TS, so no slowdown there.
ID: 25533 · Report as offensive     Reply Quote
Profileold_user81594

Send message
Joined: 11 Jun 05
Posts: 67
Credit: 1,222,916
RAC: 0
Message 26120 - Posted: 17 Jan 2007, 18:27:00 UTC - in response to Message 25530.  

These 1.10-1.30 speeds people report are overclocked machines, and the beauty about this chip is that they are rather easy to overclock.

Mine E6600 @3.2 GHz is doing 1.32-1.35 with two models and the Linux client (which is a couple of % slower than Windows)
With one model it\'s around 1.25.
With one model + a GIMPS M32 LL test it\'s around 1.45, hmm...



Hi,

Re: Time-Step speeds.

I think you are right. I have just installed a new SATA II hard-drive, and re-installed XP and everything else. My E6600 is currently overclocked to a very stable 3GHz (333MHz FSB) @ 667 DDR2 RAM. I have on both cores 1.35s/TS.
This has dropped from 1.62s/TS on a stock 2.4GHz E6600 with 800MHz RAM. The maths says that this is completely right:

3/2.4 = 1.25 (ie. 25% quicker CPU clock)

1.62 secs. / 1.25 = 1.3 secs.

I\'m getting 1.32 which is only a smidge slower, probably due to the slower 667MHZ RAM clock as opposed to 800MHz.

Neil.
ID: 26120 · Report as offensive     Reply Quote
Profileold_user202885
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Oct 06
Posts: 13
Credit: 8,437
RAC: 0
Message 26130 - Posted: 17 Jan 2007, 21:12:18 UTC - in response to Message 26120.  

we\'re beta testing a new version of the hadcm3 model, using the latest Intel Fortran compiler (9.1) and some optimizations which seem to work. People are getting \"sub-1-second\" per timesteps on Core2 Duos (these CPUs seem to be the best; I guess Intel has made their compiler work best on their latest & greatest chip! :-)

ID: 26130 · Report as offensive     Reply Quote
Desti

Send message
Joined: 6 Aug 04
Posts: 124
Credit: 9,195,838
RAC: 0
Message 26285 - Posted: 22 Jan 2007, 22:43:58 UTC - in response to Message 26130.  

we\'re beta testing a new version of the hadcm3 model, using the latest Intel Fortran compiler (9.1) and some optimizations which seem to work. People are getting \"sub-1-second\" per timesteps on Core2 Duos (these CPUs seem to be the best; I guess Intel has made their compiler work best on their latest & greatest chip! :-)



WoW! I have the new version will be ready when my current model is finished. :-)
Linux Users Everywhere @ BOINC
ID: 26285 · Report as offensive     Reply Quote
Profileold_user215584
Avatar

Send message
Joined: 26 Dec 06
Posts: 23
Credit: 11,431
RAC: 0
Message 26297 - Posted: 23 Jan 2007, 4:11:09 UTC

I have a C2D E6400 ready and waiting for somethings it would work well at. Until then ill have my 1.9-2.1secs/ts i get now :(
ID: 26297 · Report as offensive     Reply Quote
Profileold_user81594

Send message
Joined: 11 Jun 05
Posts: 67
Credit: 1,222,916
RAC: 0
Message 26495 - Posted: 1 Feb 2007, 18:17:12 UTC - in response to Message 26297.  

I have a C2D E6400 ready and waiting for somethings it would work well at. Until then ill have my 1.9-2.1secs/ts i get now :(



I think 1.9s/TS is about right for your PC, but not 2.1.

My stock E6600 is 1.62s/TS (2.4GHz), so yours with 13% slower clock gives 1.82s/TS and then knock say 3% off for having only half the cache (guesstimate figure....) comes to 1.88s/TS.

Neil.
ID: 26495 · Report as offensive     Reply Quote
ProfilePooh Bear 27
Avatar

Send message
Joined: 5 Feb 05
Posts: 465
Credit: 1,914,189
RAC: 0
Message 26496 - Posted: 1 Feb 2007, 18:22:45 UTC

My stock E6400 @ 2.13G is getting about 1.75t/s.
ID: 26496 · Report as offensive     Reply Quote
Profileold_user215584
Avatar

Send message
Joined: 26 Dec 06
Posts: 23
Credit: 11,431
RAC: 0
Message 26502 - Posted: 2 Feb 2007, 5:31:38 UTC

I am on Linux and im told the Linux app is slower so that probably explains the difference away.
ID: 26502 · Report as offensive     Reply Quote
Profilegeophi
Volunteer moderator

Send message
Joined: 7 Aug 04
Posts: 2187
Credit: 64,822,615
RAC: 5,275
Message 26522 - Posted: 2 Feb 2007, 19:10:05 UTC - in response to Message 26502.  

I am on Linux and im told the Linux app is slower so that probably explains the difference away.

Yeah, Linux is a little slower, but not a huge amount. My FX57 gets 1.46 s/TS in Windows and 1.52 s/TS in Linux. I\'m wondering if your performance has something to do with the memory configuration...whether it\'s running dual channel or not. What motherboard do you have?
ID: 26522 · Report as offensive     Reply Quote
Profileold_user215584
Avatar

Send message
Joined: 26 Dec 06
Posts: 23
Credit: 11,431
RAC: 0
Message 26553 - Posted: 3 Feb 2007, 12:28:24 UTC

Hmm well i only have 1Gb RAM of Dual channel corsair 675Mhz ram on an Intel 965 chipset (Intel motherboard) with onboard video. it shows up as close to 1Gb in Ksysguard so i don\'t think the Vid card is poaching too much RAM.
ID: 26553 · Report as offensive     Reply Quote

Message boards : Number crunching : Core 2 Duo E6600 speed

©2024 cpdn.org