Core i7 running Seti

Message boards : Number crunching : Core i7 running Seti
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 . . . 11 · Next

AuthorMessage
archae86

Send message
Joined: 31 Aug 99
Posts: 909
Credit: 1,582,816
RAC: 0
United States
Message 826603 - Posted: 3 Nov 2008, 17:05:38 UTC - in response to Message 826601.  

Well It's not that obvious, I mean Neh_1 and Neh_2 could have been called:

Neh_64

Neh_32

More obvious isn't It?
Yes, it would have been more obvious regarding that point. At the time I chose the names I had no idea the hosts differed in this respect. They may also differ in others.

By the way, thanks for deleting the images from your initial quoted reply, I appreciate it.

ID: 826603 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65773
Credit: 55,293,173
RAC: 49
United States
Message 826608 - Posted: 3 Nov 2008, 17:24:20 UTC - in response to Message 826603.  

Well It's not that obvious, I mean Neh_1 and Neh_2 could have been called:

Neh_64

Neh_32

More obvious isn't It?
Yes, it would have been more obvious regarding that point. At the time I chose the names I had no idea the hosts differed in this respect. They may also differ in others.

By the way, thanks for deleting the images from your initial quoted reply, I appreciate it.

Not a problem, Although I could have shrunk It using "th_" added on to the name of the file with a link to the full sized one over at photobucket.
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 826608 · Report as offensive
Profile Francois Piednoel
Avatar

Send message
Joined: 14 Jun 00
Posts: 898
Credit: 5,969,361
RAC: 0
United States
Message 826618 - Posted: 3 Nov 2008, 17:49:03 UTC - in response to Message 826585.  
Last modified: 3 Nov 2008, 17:50:07 UTC



Key questions:

1. what actual clock frequency are the Nehalem hosts running (ap
reporting appears to be unreliable, though the central tendency may be
toward a slight overclock).
2. were the Nehalem results generated running hyperthreaded?
3. how many SETI aps were actually running simutaneously when these
were generated?
4. what fraction of wall clock hours since these two hosts were signed
up to SETI have they actually been running SETI? The current RAC
seems lower than my optimistic assessment of their capabilities if
runnning 8 aps 24x7, especially for the longer-running host 4603078
(Neh_2).



1) my Nhm is running at 4.0Ghz

2) Yes, HT is on, and awesome :-P

3) 8 at the same time.

4) my Core i7 is running seti since 9th of October .. i have been hiding ;-)

it did run 24/7, during the tunning for Overclocking, few half day down due to tunning.

I expect the RAC to get around 9500, am I right?
ID: 826618 · Report as offensive
archae86

Send message
Joined: 31 Aug 99
Posts: 909
Credit: 1,582,816
RAC: 0
United States
Message 826627 - Posted: 3 Nov 2008, 18:39:38 UTC - in response to Message 826618.  

1) my Nhm is running at 4.0Ghz

2) Yes, HT is on, and awesome :-P

3) 8 at the same time.

4) my Core i7 is running seti since 9th of October .. i have been hiding ;-)

it did run 24/7, during the tunning for Overclocking, few half day down due to tunning.

I expect the RAC to get around 9500, am I right?

Thanks for the answers, Francois.

The one I labelled Neh_1 is the one linked in the opening message of this thread, so we can take it that most of the points for it represent 4 GHz operation. That is a 25% overclock to apparent spec, where my Q6600 and E6600 are at 16.25% overclock, and my Q9550 right at stock. Those wishing to compare stock speeds will need to make appropriate adjustments. Those wishing to compare a given degree of overclock difficulty probably need to wait.

Adjustments aside, the total capability for these two hosts in a four-core CPU is obviously very large. When reasonably priced motherboards and RAM become available, this may become an attractive option for serious SETI crunchers inhibited by a value orientation. I'd think Mark Sattler eager to jump sooner.

I don't have much interest in predicting RAC, but as you gave such crisp and prompt answers I feel obligated to take a try. It appears than in VLAR and mid AR your mean cost is about 65 seconds/credit. On the false assumption of 100% "CPU efficienty" as BOINC defines it that would allow an eventual RAC slightly over 10500. Any services, downtime, and so forth lowering the resource available to SETI would reduce that, though I'd think beating 10,000 is thinkable. Of course a shift in the character of the WUs being distributed (both noise level and AR) moves things up and down as well.

So on what I see your 9500 seems a realistic goal at your current operating point.
ID: 826627 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51469
Credit: 1,018,363,574
RAC: 1,004
United States
Message 826628 - Posted: 3 Nov 2008, 18:40:44 UTC - in response to Message 826618.  


I expect the RAC to get around 9500, am I right?

I would think that or even better......
The Frozen Penny has 4 cores running at 4.4ghz and has reached 8500 RAC....

I should think that 8 threads crunching at 4ghz would do better than 9500....
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 826628 · Report as offensive
Profile Paul D Harris
Volunteer tester

Send message
Joined: 1 Dec 99
Posts: 1122
Credit: 33,600,005
RAC: 0
United States
Message 826630 - Posted: 3 Nov 2008, 18:54:46 UTC

If it weren’t for the Hyper-threading the Corei7 would be a big yawn. The Hyper-threading is it saving grace. Because it is just as fast as my Q6600 at stock speed it will do a 17 credit Wu in 1200 seconds the same as the Corei7. So Hyper-threading doubles the workload.
ID: 826630 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51469
Credit: 1,018,363,574
RAC: 1,004
United States
Message 826631 - Posted: 3 Nov 2008, 18:56:24 UTC - in response to Message 826630.  

If it weren’t for the Hyper-threading the Corei7 would be a big yawn. The Hyper-threading is it saving grace. Because it is just as fast as my Q6600 at stock speed it will do a 17 credit Wu in 1200 seconds the same as the Corei7. So Hyper-threading doubles the workload.

Dunno.....but the Frozen Penny does 17.XX credit WUs in about 495 seconds........
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 826631 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65773
Credit: 55,293,173
RAC: 49
United States
Message 826636 - Posted: 3 Nov 2008, 19:10:49 UTC - in response to Message 826628.  


I expect the RAC to get around 9500, am I right?

I would think that or even better......
The Frozen Penny has 4 cores running at 4.4ghz and has reached 8500 RAC....

I should think that 8 threads crunching at 4ghz would do better than 9500....

Mark I think the Frozen Penny has been drawn and quartered by the i7. ;)
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 826636 · Report as offensive
archae86

Send message
Joined: 31 Aug 99
Posts: 909
Credit: 1,582,816
RAC: 0
United States
Message 826637 - Posted: 3 Nov 2008, 19:10:55 UTC - in response to Message 826630.  

If it weren’t for the Hyper-threading the Corei7 would be a big yawn. The Hyper-threading is it saving grace. Because it is just as fast as my Q6600 at stock speed it will do a 17 credit Wu in 1200 seconds the same as the Corei7. So Hyper-threading doubles the workload.
Neither of us knows how fast the cores would be without hyperthreading unless you have a source I've not seen. The early reviews were done on boards/BIOSes that could not turn it off. I don't know whether one will be able to turn it off on production boards.

I did some studies of the hypthreading benefit on my Northwood-descended P4. It depended on ap and such, but a typical number which sticks in my brain from then is about 20%. In other words, with hyperthreading turned off the CPU seconds reported for a WU dropped drastically.

Of course, the better the implementation of HT fits a specific aps, the less the time would drop--but your comments suggest you assume it would not drop at all, which strikes me as very unlikely.

ID: 826637 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51469
Credit: 1,018,363,574
RAC: 1,004
United States
Message 826640 - Posted: 3 Nov 2008, 19:16:58 UTC - in response to Message 826636.  


I expect the RAC to get around 9500, am I right?

I would think that or even better......
The Frozen Penny has 4 cores running at 4.4ghz and has reached 8500 RAC....

I should think that 8 threads crunching at 4ghz would do better than 9500....

Mark I think the Frozen Penny has been drawn and quartered by the i7. ;)

Not per core it hasn't...........
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 826640 · Report as offensive
archae86

Send message
Joined: 31 Aug 99
Posts: 909
Credit: 1,582,816
RAC: 0
United States
Message 826641 - Posted: 3 Nov 2008, 19:17:31 UTC - in response to Message 826631.  

Dunno.....but the Frozen Penny does 17.XX credit WUs in about 495 seconds........
One little tidbit, which illustrates the point that the Nehalem is meaningfully different architecturally is that the VHAR stuff we are used to being a rich source of credit rate is actually its worst operating region of those available for review. This is strongly true for the 64-bit build 42 stuff--and even seems to get gradually worse as AR rises in this region.

Since VHAR makes up a pretty low proportion of the total workload, weighted either by CPU seconds or credit, it is poor region on which to base overall comparison. Cherry-pickers please ignore this discussion.

ID: 826641 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51469
Credit: 1,018,363,574
RAC: 1,004
United States
Message 826643 - Posted: 3 Nov 2008, 19:21:17 UTC - in response to Message 826641.  

Dunno.....but the Frozen Penny does 17.XX credit WUs in about 495 seconds........
One little tidbit, which illustrates the point that the Nehalem is meaningfully different architecturally is that the VHAR stuff we are used to being a rich source of credit rate is actually its worst operating region of those available for review. This is strongly true for the 64-bit build 42 stuff--and even seems to get gradually worse as AR rises in this region.

Since VHAR makes up a pretty low proportion of the total workload, weighted either by CPU seconds or credit, it is poor region on which to base overall comparison. Cherry-pickers please ignore this discussion.

OK.....for the sake of the argument......here are some other Frozen Penny timings.........
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 826643 · Report as offensive
Profile Sir Ulli
Volunteer tester
Avatar

Send message
Joined: 21 Oct 99
Posts: 2246
Credit: 6,136,250
RAC: 0
Germany
Message 826667 - Posted: 3 Nov 2008, 20:20:01 UTC

8 CPUs that is great i think...

Greetings from Germany NRW
Ulli


ID: 826667 · Report as offensive
Profile Gecko
Volunteer tester
Avatar

Send message
Joined: 17 Nov 99
Posts: 454
Credit: 6,946,910
RAC: 47
United States
Message 826668 - Posted: 3 Nov 2008, 20:22:00 UTC

Got some apples and oranges being compared in this discussion.
The playing field that Mark's frozen penny and Corei7 are on are quite a bit different & w/ different rules.

However, one thing appears pretty clear.
A person can get the Extreme chip & do a pretty simple OC w/ Skulltrail results....for @ 1/2 or less of the cost & I'd wager that one could build one for less than the frozen penny.

Or...just wait for new Mac Pro which is likely to debut the dual slot boards in a few months.
If they manage 16 cores for @ 3K...I'll be pretty tempted to update the "family pc" ; > )



ID: 826668 · Report as offensive
Profile Paul D Harris
Volunteer tester

Send message
Joined: 1 Dec 99
Posts: 1122
Credit: 33,600,005
RAC: 0
United States
Message 826673 - Posted: 3 Nov 2008, 20:34:06 UTC
Last modified: 3 Nov 2008, 20:38:16 UTC

Hey Corei7
Give me a call when your cousin Octo shows up a true 8 core CPU and not a 4 core CPU that acts like an eight core CPU.
ID: 826673 · Report as offensive
Profile gizbar
Avatar

Send message
Joined: 7 Jan 01
Posts: 586
Credit: 21,087,774
RAC: 0
United Kingdom
Message 826681 - Posted: 3 Nov 2008, 21:04:18 UTC

OK! ENOUGH ALREADY!!!

Yes, it does make my 9850 phenom look like an abacus!
Yes, I am gonna have to get my calculator out to help out!
What sort of RAC am I gonna get from a Casio?

@msattler - Are we gonna get to see a frozen core i7 anytime soon? Or is the piggy bank empty after the Seagate Drives fiasco?

regards, Gizbar.



A proud GPU User Server Donor!
ID: 826681 · Report as offensive
OzzFan Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Apr 02
Posts: 15691
Credit: 84,761,841
RAC: 28
United States
Message 826707 - Posted: 3 Nov 2008, 21:52:07 UTC

The Core i7 looks like a fairly exciting product. I just might have to build myself one come tax time. ;)
ID: 826707 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51469
Credit: 1,018,363,574
RAC: 1,004
United States
Message 826719 - Posted: 3 Nov 2008, 22:04:59 UTC - in response to Message 826681.  

OK! ENOUGH ALREADY!!!

Yes, it does make my 9850 phenom look like an abacus!
Yes, I am gonna have to get my calculator out to help out!
What sort of RAC am I gonna get from a Casio?

@msattler - Are we gonna get to see a frozen core i7 anytime soon? Or is the piggy bank empty after the Seagate Drives fiasco?

regards, Gizbar.

Piggy bank is pretty light these days........
The next few months are pretty rough......
Income taxes, property taxes, home and auto insurance...all hit in the coming months.......
It would be a little easier if the upgrade did not require a new mobo and 3 matched sticks of RAM.....
But I cannot complain really......the ol' socket 775 has been a real workhorse.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 826719 · Report as offensive
Profile Fuzzy Hollynoodles
Volunteer tester
Avatar

Send message
Joined: 3 Apr 99
Posts: 9659
Credit: 251,998
RAC: 0
Message 826730 - Posted: 3 Nov 2008, 22:16:47 UTC

From Danish PC World, translated in Google Translate:

Test: So fast is Intel's new Core processor i7

Sorry for the bad translation here and there. I have submitted more accurate translations where it's not what it should be...


"I'm trying to maintain a shred of dignity in this world." - Me

ID: 826730 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65773
Credit: 55,293,173
RAC: 49
United States
Message 826758 - Posted: 3 Nov 2008, 23:18:34 UTC - in response to Message 826681.  

OK! ENOUGH ALREADY!!!

Yes, it does make my 9850 phenom look like an abacus!
Yes, I am gonna have to get my calculator out to help out!
What sort of RAC am I gonna get from a Casio?

@msattler - Are we gonna get to see a frozen core i7 anytime soon? Or is the piggy bank empty after the Seagate Drives fiasco?

regards, Gizbar.

Don't feel bad Gizbar, Even My cpus are a bit left behind to one extent or another, But suffice It to say, You might have an Abacus there, But It's a very fast one, Faster than My Opteron 165 is I'd think(Bronze age tool). :D
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 826758 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 . . . 11 · Next

Message boards : Number crunching : Core i7 running Seti


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.