so that's a "no" to the data comparison with actual facts?

I really am curious. I see benchmark result of the E6600, E6700, E6800, and QX6700. I'm curious about how they really compare. I suspect my AMD64 5200 falls between the E6700 and E6800, and would like to know. I've even been looking at the Asus 1LN64 mobo with two FX74 chips, although I think they're a bit (ok alot)Pricey.

OK, I'll tell you what, you keep your computers visible and provide me with your watt VA, and operating voltage readings for a "no load" and "loaded (all 8 cores)", and I'll gather the data myself. I'll do the heavy lifting.

tony

Processing power density is a very big factor for crunchers, your 8 Boxes will be 8 times bigger ... Processing power wise, you may get a little a head, because you paid for the extra memory ... price wise, you ll get about the same. POWER wise, you ll have 8 hard drives ... 8 GPUs, or intergrated graphics ... what do you think? From a watt point of view, you ll be way a head, using much more power! your processing power density will be very bad, and your Performance per watt will be down to the toillette, because the HDD and other additional devices.

Most of the big crunching farms tried this and they came back from it. I don t even speak about the faillure of your system, you multiply the chance of faillure by 8!

you can do your own mathematics, but many expert did before, it is not a good alternative.

who?

Processing power density is a very big factor for crunchers, your 8 Boxes will be 8 times bigger ... Processing power wise, you may get a little a head, because you paid for the extra memory ... price wise, you ll get about the same.

same what? same overall cost/decoy? Same credit? ??????

POWER wise, you ll have 8 hard drives ... 8 GPUs, or intergrated graphics ... what do you think? From a watt point of view, you ll be way a head, using much more power! your processing power density will be very bad, and your Performance per watt will be down to the toillette, because the HDD and other additional devices.

I agree 8 of mine will draw more, or atleast I think they will. What I endeavor to discover is a comparable cost/decoy (on simlar wus) so I can take that into account.

Most of the big crunching farms tried this and they came back from it. I don t even speak about the faillure of your system, you multiply the chance of faillure by 8!

I concur. Also, if I have a hard drive failure, 7/8th of my farm is still running. Yours' would be down.

you can do your own mathematics, but many expert did before, it is not a good alternative.

who?

that's what I would like to do, but can't without the figures to input into the formulas.

Just to add a monkey-wrentch (?) into all of this, Penryn to have HT

Quad Penryn is still just two dies stuck together in the package! :-( Had been hoping for something a bit more sophisticated. Perhaps a better quad is what is coming in August.

http://www.reghardware.co.uk/2007/01/27/intel_reveals_penryn/

On the other hand, there are process improvements which the Inq explains better than the Reg:
http://www.theinquirer.net/default.aspx?article=37253

---

Processing power density is a very big factor for crunchers, your 8 Boxes will be 8 times bigger ... Processing power wise, you may get a little a head, because you paid for the extra memory ... price wise, you ll get about the same. POWER wise, you ll have 8 hard drives ... 8 GPUs, or intergrated graphics ... what do you think? From a watt point of view, you ll be way a head, using much more power! your processing power density will be very bad, and your Performance per watt will be down to the toillette, because the HDD and other additional devices.

This is very true - to mitigate this I run mine from compactflash using the XPe enhanced write filter - even my backup server which has a HD is going to boot from a 1GB CF card and then the HD will only spin up when backing up data. I recommend it to anyone!

There are still more PSUs and northbridge chips etc...

---

Quad Penryn is still just two dies stuck together in the package! :-( Had been hoping for something a bit more sophisticated. Perhaps a better quad is what is coming in August.

ok, so, why do you want it to be on 1 die, only company that have nothing else to sell do that. The native quad core is a bad idea, as long as you can not produce them with a high rate of success, or you are ready to produce them as a lost for marketing purpose.
when your process of fabrication is not tuned enough, producing a huge die like a quad core is losing money...
good articcle about process

on the top of this, it is still to be proven that a native quad core can beat a non native one, AMD is doing a lot of claim, but it is only on paper ... and on 1 benchmark out of 200 ... i can bid a lot of money on the fact that the native quad core will not beat non native quad Core 2 QX6700 on many benchmarks, such as DivX, cinebench, 3Dsmax, povray, ROSETTA, SETI ....

it is a marketing hype that AMD used before, the only difference is this time, they ll get distroy on the benchmark suites ...

by the way, the multipackaging system used in the Quad from intel is much more advanced than regular technics, if AMD was able to do it, trust me, they will!
they have almost 2 quaters without quad cores, they are so dessesperated that they had to create a dual socket motherboard to compete with the Quads. the result was dissapointing, and the result is not going to be better with barcelona. AMD is trying to keep their fan boys happy, but on paper, this is ok ... when it gets to silicon, they are now one generation behind, architecture wise and process wise, not really a good position to be when you claim a 20% in 2007

The race is over.

who?

this is my own opinion.

Most of the big crunching farms tried this and they came back from it. I don t even speak about the faillure of your system, you multiply the chance of faillure by 8!

I concur. Also, if I have a hard drive failure, 7/8th of my farm is still running. Yours' would be down.

But yours is going to fail statistically 8 times more, giving you 8 times longer down times. and costing you 8 times more to repare.

This is a non winning proposition for professionals... if you like to do it, do it, but at least, use a Fast CPU for good price:
Can't beat this one for what you want to do: Core 2 E4300!!!
I quote: "Every generation a ridiculously overclockable bargain chip is born, and the Core 2 Duo E4300 is just that chip. Although it's not yet launched, Intel has a true winner on its hands with the E4300."

who?

For crying in a bucket. I started collecting the data from Whos?s' monster. I started at 11 a.m. I sorted them by resultID (date) and started at the top (18 Jan), and I was going along just fine (extracting wu name and decoy count)then about the 22nd he had a 100 or so client errors. "Ok, I'll just delete them", I said to myself. On the 18th he was using a 6 hr prefs, and a few 1 hours mixed in, then on the 19th he switched to two hours (Man, there's a buttload of them at 2 hr pref), but I kept chugging. Then on the 24th I thought everything would be good, he's switched to 24 hour run time pref (man this will be fast now), then after ONE day he switched it to ONE HOUR. Yikes, that's lots of work units to wade through. He still has this setting, I've spent 6 hours scrapping so far(decent nap and tv viewing included) and think I'll stop after the 24 of Jan (issue date).

Who? I have part of my answer. Do you care if I display the work done by your V8 to these boards? There's no data that isn't already available, but I thought I'd ask.

tony

NOTE: it won't be today/tonite. I just have the raw data for both yours' and mine extracted. I'm not up to staring at it any more today.

Who? I have part of my answer. Do you care if I display the work done by your V8 to these boards? There's no data that isn't already available, but I thought I'd ask.

tony

NOTE: it won't be today/tonite. I just have the raw data for both yours' and mine extracted. I'm not up to staring at it any more today.

we all know the answer ... you ll get more credit done, if you use the new core 2 E series, you should get about double for the same budgets, but again, you are comparing Honda Civics and Lambo Diablo ... you can drive more people ... but the ladies come with me :-P (hehehe kidding)


who?

I started collecting the data from Whos?s' monster.

If you would like another machine to collect data on, here's an interesting one:

http://setiathome.berkeley.edu/show_host_detail.php?hostid=3022805

It is a Mac V8. =;^) Basically a Mac Pro that had the 2x 51xx replaced with 2x E5345. Oh yea, he also has alexkan's smokin' app. It's been generating over 7000 per day, and has a RAC of over 4500 after only *9* days of crunching.

FYI: The apple store has refurb'd 2x2.0ghz Mac Pros in stock for $1899. Pretty good way to go, assuming you are going to upgrade to dual quads. Add another $950 each for the E5345s, and another 2x 512gb FB DIMMs for about $220 (need 4x DIMMs for best performance), and you have a fast 8-way Mac Pro for about $4100 + Tax. Not bad....

---

Reno, NV
Team: SETI.USA

I started collecting the data from Whos?s' monster.

If you would like another machine to collect data on, here's an interesting one:

http://setiathome.berkeley.edu/show_host_detail.php?hostid=3022805

It is a Mac V8. =;^) Basically a Mac Pro that had the 2x 51xx replaced with 2x E5345. Oh yea, he also has alexkan's smokin' app. It's been generating over 7000 per day, and has a RAC of over 4500 after only *9* days of crunching.

FYI: The apple store has refurb'd 2x2.0ghz Mac Pros in stock for $1899. Pretty good way to go, assuming you are going to upgrade to dual quads. Add another $950 each for the E5345s, and another 2x 512gb FB DIMMs for about $220 (need 4x DIMMs for best performance), and you have a fast 8-way Mac Pro for about $4100 + Tax. Not bad....

be patient ... as soon as I get autorized to use SSE4, I ll show you ...
get ready for "wroooommmmmmmmmmmmmmmm"

who?

be patient ... as soon as I get autorized to use SSE4, I ll show you ...
get ready for "wroooommmmmmmmmmmmmmmm"

When it comes to volunteer work like this, it's bad form to push too hard. I'll let it slide for now...=;^)

Seriously though. Which chips will be able to work with:

SSSE3
SSE4

?

---

Reno, NV
Team: SETI.USA

be patient ... as soon as I get autorized to use SSE4, I ll show you ...
get ready for "wroooommmmmmmmmmmmmmmm"

When it comes to volunteer work like this, it's bad form to push too hard. I'll let it slide for now...=;^)

Seriously though. Which chips will be able to work with:

SSSE3
SSE4

?

More detail here

Penryn and 45nm Co will have SSE4, it means, you ll get it in 2007 :)

can 't say more except that it rocks!


who?

Notice that my both machines are running VISTA ultimate :)


who?

who?

i just read this which mentions that there will be a return of hyperthreading in the wolfdale core, but without doubling the virtual cores as it has always involved in the P4s. Does this mean that each thread will run on two cores (so 2 cores would run 2 threads, but with neither thread specifically assigned to either core), to make the most of the spare cycles when there are latencies?

Can you comment?

cheers
Danny

---

who?

i just read this which mentions that there will be a return of hyperthreading in the wolfdale core, but without doubling the virtual cores as it has always involved in the P4s. Does this mean that each thread will run on two cores (so 2 cores would run 2 threads, but with neither thread specifically assigned to either core), to make the most of the spare cycles when there are latencies?

Can you comment?

cheers
Danny

I can not comment on any Intel futur product, this is trade secret, and I should not say anything. I can comment only on actual intel product, and un-officially, because I love my job, and i don t want to lose it :)

who?

who?

I have been following this thread very closely, and I would like to ask you a couple of comparison questions. I do not want any info you don't feel comfortable in offering. I would be happy to post an e-mail address to reach me if you rather not have the question asked here. Or you can simply decline and no worries.

sirslacker

---

... all about the work ...
... time to step-up and throw-down ...

who?

I have been following this thread very closely, and I would like to ask you a couple of comparison questions. I do not want any info you don't feel comfortable in offering. I would be happy to post an e-mail address to reach me if you rather not have the question asked here. Or you can simply decline and no worries.

sirslacker

For comparaisons, you have Anandtech.com ... I am not here to do marketing for my employer.
I am here to have fun, without giving away any of my employer intellectual property and secrets.
If you need performance informations, in a business matter, i ll be happy to connect you with the right people.

When I speak about SSE4, it is because the intruction set is public, i am very carefull to do not give away any strategic information, here or by email.

who?

who?

I have been following this thread very closely, and I would like to ask you a couple of comparison questions. I do not want any info you don't feel comfortable in offering. I would be happy to post an e-mail address to reach me if you rather not have the question asked here. Or you can simply decline and no worries.

sirslacker

For comparaisons, you have Anandtech.com ... I am not here to do marketing for my employer.
I am here to have fun, without giving away any of my employer intellectual property and secrets.
If you need performance informations, in a business matter, i ll be happy to connect you with the right people.

When I speak about SSE4, it is because the intruction set is public, i am very carefull to do not give away any strategic information, here or by email.

who?

Anyone know when AMD will incorporate SSSE3 and SSE4 or some 3DNow! Pro Extra Ultimate into their chips ?

---

Team mauisun.org