V8 / AMD64 X2 5200 WU comparison

Message boards : Number crunching : V8 / AMD64 X2 5200 WU comparison

To post messages, you must log in.

AuthorMessage
Astro
Avatar

Send message
Joined: 2 Oct 05
Posts: 987
Credit: 500,253
RAC: 0
Message 35670 - Posted: 28 Jan 2007, 13:23:19 UTC
Last modified: 28 Jan 2007, 13:25:26 UTC

Collection period was all results issued Jan 18 to Jan 24 for both machines. Decoy and Result Description were extracted from the result ID of each and added to the spreadsheet. The data for the V8 is in Green, and the X2 5200 is in red. Data compiled then sorted by "result description". Where "result description" seemed to match up for both computers, data was kept, where there was no match, data was discarded. (trying to compare apples to apples).

Analysis to follow.

I'm not a whiz at rosetta result identities so if anyone sees where I'm keeping the wrong ones (not comparing apples to apples), please let me know.

NOTE: CPU time varies




ID: 35670 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Who?

Send message
Joined: 2 Apr 06
Posts: 213
Credit: 1,366,981
RAC: 0
Message 35677 - Posted: 28 Jan 2007, 17:37:01 UTC - in response to Message 35670.  


ok ... so?

who?
ID: 35677 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
BennyRop

Send message
Joined: 17 Dec 05
Posts: 555
Credit: 140,800
RAC: 0
Message 35689 - Posted: 29 Jan 2007, 0:06:12 UTC

It shows that you'd need about 50% more AMD X2 5200 cores to match the performance of the unavailable Intel part being shown. We can then decide if hordes of cheaper X2 5200s are better for us than waiting for the release of the new Intel part.
ID: 35689 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Astro
Avatar

Send message
Joined: 2 Oct 05
Posts: 987
Credit: 500,253
RAC: 0
Message 35690 - Posted: 29 Jan 2007, 3:08:09 UTC
Last modified: 29 Jan 2007, 3:08:49 UTC

Sorry, I only had this AM to devote to this, and no time tonite. I think it's interesting to look at the "granted credit/decoy" column and see that it (the new credit system) really seems to work. Also looking at granted credit/hour. 15.XX is just about right for my X2, and can see that his "seconds/decoy" being X amount faster than my X2 nets him the same X amount increase in granted credit/hour (roughly 20/hour).

as I stated earlier, "analysis to follow". that means totals/averages/etc. Just thought someone might find it interesting, since although the data is available, it's not presented in this fashion.
ID: 35690 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Who?

Send message
Joined: 2 Apr 06
Posts: 213
Credit: 1,366,981
RAC: 0
Message 35691 - Posted: 29 Jan 2007, 3:35:54 UTC - in response to Message 35689.  
Last modified: 29 Jan 2007, 3:36:27 UTC

It shows that you'd need about 50% more AMD X2 5200 cores to match the performance of the unavailable Intel part being shown. We can then decide if hordes of cheaper X2 5200s are better for us than waiting for the release of the new Intel part.


I don't know how you did your math ... but you are wrong.
those big map are useless, an XLS files will be better. You ll need more than 4 times for sure, based on few addition of 20 lines I did ...
50% is in your dreams ...

I don t see the 24hours work load in this bitmap eather, so, back to the counting ... many parts are missing.

you are comparing apple and banana , what matter is granded credit, and only this!
As i expected, it is useless.



who?
ID: 35691 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Astro
Avatar

Send message
Joined: 2 Oct 05
Posts: 987
Credit: 500,253
RAC: 0
Message 35693 - Posted: 29 Jan 2007, 3:43:46 UTC

I'll put my sig here, so everyone will have an link to the "Boinc help desk", from there, you can pick my name and email me. I'll send you a copy of the "full" xls file. This shows every wu I've done (since recording started), and even has a worksheet labelled "who". Showing all his work from Jan 18 to Jan 24. "sheet3" worksheet is where I took the pics to get the one below. Note: the work sheets are labelled with my host name and if linux was the OS then a linux tag was added, so there's two worksheets/host (where I used two oses to do work).

The who worksheet is much larger than what I showed below. Where each machine did what appears to be the same style wu, I kept it. Where there was no "commom wu" I discarded them.

nite all
ID: 35693 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Astro
Avatar

Send message
Joined: 2 Oct 05
Posts: 987
Credit: 500,253
RAC: 0
Message 35694 - Posted: 29 Jan 2007, 3:44:10 UTC

darn, I'm so used to unchecking the sig box, sorry


ID: 35694 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
BennyRop

Send message
Joined: 17 Dec 05
Posts: 555
Credit: 140,800
RAC: 0
Message 35698 - Posted: 29 Jan 2007, 7:05:14 UTC - in response to Message 35691.  

It shows that you'd need about 50% more AMD X2 5200 cores to match the performance of the unavailable Intel part being shown. We can then decide if hordes of cheaper X2 5200s are better for us than waiting for the release of the new Intel part.


I don't know how you did your math ... but you are wrong.
those big map are useless, an XLS files will be better. You ll need more than 4 times for sure, based on few addition of 20 lines I did ...
50% is in your dreams ...

I don t see the 24hours work load in this bitmap eather, so, back to the counting ... many parts are missing.

you are comparing apple and banana , what matter is granded credit, and only this!
As i expected, it is useless.[/quote]

The v8 used around 200 sec/decoy; the X2 5200 used around 300 sec/decoy for a similar set of WUs. (listed with 1.42 and 1.43 credits/decoy). Other WUs showed a higher ratio, while others showed a lesser ratio that ranged between 1 and 2. It seems to hover around the 1.5 ratio - or a 50% increase. Thus 3 x2 5200 cores should equal 2 of your v8 cores under Rosetta. Or 12 x2 5200 cores to roughly equal the 8 cores.
ID: 35698 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Who?

Send message
Joined: 2 Apr 06
Posts: 213
Credit: 1,366,981
RAC: 0
Message 35699 - Posted: 29 Jan 2007, 7:36:52 UTC - in response to Message 35698.  

It shows that you'd need about 50% more AMD X2 5200 cores to match the performance of the unavailable Intel part being shown. We can then decide if hordes of cheaper X2 5200s are better for us than waiting for the release of the new Intel part.


I don't know how you did your math ... but you are wrong.
those big map are useless, an XLS files will be better. You ll need more than 4 times for sure, based on few addition of 20 lines I did ...
50% is in your dreams ...

I don t see the 24hours work load in this bitmap eather, so, back to the counting ... many parts are missing.

you are comparing apple and banana , what matter is granded credit, and only this!
As i expected, it is useless.


The v8 used around 200 sec/decoy; the X2 5200 used around 300 sec/decoy for a similar set of WUs. (listed with 1.42 and 1.43 credits/decoy). Other WUs showed a higher ratio, while others showed a lesser ratio that ranged between 1 and 2. It seems to hover around the 1.5 ratio - or a 50% increase. Thus 3 x2 5200 cores should equal 2 of your v8 cores under Rosetta. Or 12 x2 5200 cores to roughly equal the 8 cores. [/quote]


I was expecting 8 ... 12 is a good news for AMD, they can sell more slow processor ;)

who?
ID: 35699 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Astro
Avatar

Send message
Joined: 2 Oct 05
Posts: 987
Credit: 500,253
RAC: 0
Message 35709 - Posted: 29 Jan 2007, 12:43:49 UTC
Last modified: 29 Jan 2007, 13:12:32 UTC

Well, I was going to calculate the coefficient on each group and come up with hard numbers, but I'd agree the V8 does roughly 50% more work per core. I'd also agree that it looks like I'd need 6 AMDX2 5200's to roughly equal the work output of ONE V8. I was guessing this from the beginning, but now we know a number of cpus and it's 6. So, 6 X $760/machine equals $4,560 vs $5,000 for the V8.

AMD64 X2 5200 cpu $290
Asus M2NPV-VM mobo $85
Enermax noisetaker 425W psu $80
2G DDR800 ram $200
Case w/o psu $40
DVD/CDrom drive $20
[edit}forgot the hard drive} $45

So, $5000-$4560= $440 bucks saved, but is it? I have electrical use to account for. I have the current draw figures for the X2, but not for the V8. So we're just down to guessing again.

also, the X2 has a modest %5 OC, Not sure what he's done to the V8

also, what's a decent length of time for life expectancy? 3 yr, 5yr??
ID: 35709 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Astro
Avatar

Send message
Joined: 2 Oct 05
Posts: 987
Credit: 500,253
RAC: 0
Message 35712 - Posted: 29 Jan 2007, 13:00:56 UTC

My X2 5200 draws 144 watts under full load. 144x24hr=3456 watts/day or 3456x365days= 1,261,440 watts/year. 1,261,440 w/year divided by 1,000=1261.44 Kwatts/year x 13.2 cents/kwh=16651 cents/year, or $166.51/year x 6 machines= $999/year electical oost for 6 of them running 24/7. Ofcourse we need his figures to see the full picture.
ID: 35712 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Feet1st
Avatar

Send message
Joined: 30 Dec 05
Posts: 1755
Credit: 4,690,520
RAC: 0
Message 35716 - Posted: 29 Jan 2007, 15:12:53 UTC

Ahhh... 6 crunchers... $4560, 144w full load 24x7... $999/yr
1000w extra heat on a cold Winter's morning... priceless!

I think the point was that a significant portion of that power requirement is for disk drives and video adapters, and mother boards... and that with a single V8 you reduce all of those consumers of electricity and cost by 88%. So, even if the electricity per operational core were higher, the total operating cost would be reduced.

Note that in a 5yr timeframe, you're spending as much on electricity as on the hardware!
Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/
ID: 35716 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Who?

Send message
Joined: 2 Apr 06
Posts: 213
Credit: 1,366,981
RAC: 0
Message 35719 - Posted: 29 Jan 2007, 16:40:16 UTC - in response to Message 35716.  

Ahhh... 6 crunchers... $4560, 144w full load 24x7... $999/yr
1000w extra heat on a cold Winter's morning... priceless!

I think the point was that a significant portion of that power requirement is for disk drives and video adapters, and mother boards... and that with a single V8 you reduce all of those consumers of electricity and cost by 88%. So, even if the electricity per operational core were higher, the total operating cost would be reduced.

Note that in a 5yr timeframe, you're spending as much on electricity as on the hardware!


so, let s try to do some AMD like marketing:

"V8 save the earth, it is very efficent against global warming"

just kidding... what ever :)

V8 is a Lambo of the crunching ... and what matter is how long is your commut, not how much CO2 you are saving in your car. I see people commuting 60 miles every day with a prius, they still emit more CO2 than me with my 4runner communiting 3 miles.

that was "the thinking of the day"

who?
ID: 35719 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Astro
Avatar

Send message
Joined: 2 Oct 05
Posts: 987
Credit: 500,253
RAC: 0
Message 35725 - Posted: 29 Jan 2007, 20:04:21 UTC

OK I went and calculated the percentage of my "Seconds/decoy" to that of the V8. Here's how it lays out per "wu group". This list is in the same order as those of my first post.



So, The percentage is not as previously guessed (66%), but rather (75%), So I need 4 cores for every 3 of his, instead of 3 for every 2, or 10.6 cores. I'm not sure how to buy the .6 core part(1/3rd of a puter), but I could the 5 and overclock a bit more. LOL

so, 5 x $760 is $3800.
ID: 35725 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Who?

Send message
Joined: 2 Apr 06
Posts: 213
Credit: 1,366,981
RAC: 0
Message 35741 - Posted: 30 Jan 2007, 3:57:36 UTC - in response to Message 35725.  
Last modified: 30 Jan 2007, 4:01:22 UTC

OK I went and calculated the percentage of my "Seconds/decoy" to that of the V8. Here's how it lays out per "wu group". This list is in the same order as those of my first post.



So, The percentage is not as previously guessed (66%), but rather (75%), So I need 4 cores for every 3 of his, instead of 3 for every 2, or 10.6 cores. I'm not sure how to buy the .6 core part(1/3rd of a puter), but I could the 5 and overclock a bit more. LOL

so, 5 x $760 is $3800.


still cost a lot of energy, and the ladies will get in my lambo ...
my other machine how many X2 do you need to beat it?
4 cores AT 4GHZ ... A REAL 4 X 4






probably about 2 OR 3 ?


who?
ID: 35741 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : V8 / AMD64 X2 5200 WU comparison



©2024 University of Washington
https://www.bakerlab.org