• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Woodcrest previewed

D

Duke3d87

Limp Gawd
Joined
Jul 6, 2004
Messages
430
IF YOU FOLLOW server and workstation processors at all, you probably have a sense that Intel's Xeon has been having some trouble competing with AMD's Opteron. Intel is looking to reverse its fortunes in the server and workstation markets with a series of new products designed to take on the Opteron and win. The first wave of those products is being introduced today in the form of a new server platform, code-named Bensley. This platform consists primarily of a new chipset that will be mated with a trio of new Xeon processors, also scheduled to arrive in stages.

The star of that lineup, once it arrives, will be the Woodcrest microprocessor based on Intel's new Core microarchitecture, which promises higher performance and lower power consumption than current Netburst-derived Xeons. I recently attended a reviewer's workshop in Portland, Oregon where I got to spend some quality time with a pair of Woodcrest processors on the Bensley platform, and I came away impressed. More importantly, I came away with some benchmark scores we can compare directly to AMD's Opteron 285.

http://www.techreport.com/onearticle.x/10021

The performance gap doesn't look as big as it did with Conroe vs. AM2 which is somewhat odd to say the least, but power consumption is very impressive. It uses less power then the Opteron 285 even running 400 MHz faster. Using AMD's powernow! thing, the Opteron used less power then the Woodcrest, but that's b/c Woodcrest didn't have like half of the power saving options as it should have. It also seems to be a very realy stepping...
 
I saw the link on /. this morning.

The 2 socket Woodcrest system tested was a little disappointing. It's still faster from the tech report's benchmarks of course, and the luckily the 5160 is $100 cheaper, runs cooler and is faster than the 285.

Maybe Poncho can give a little insight to the problem(s).
 
Woodcrest looks good in the 2 socket dual-core space. The numbers speak for themselves.

It seems two 3.0GHz DC Woodcrests should be just as fast or slightly faster than two 3.0GHz DC Socket-F Opterons.

Sounds like Intel just caught up. Maybe this will make AMD lower prices; I want to pick up a couple of faster s940 Opterons in about a year to replace my 265s.
 
Duke3d87 said:
The performance gap doesn't look as big as it did with Conroe vs. AM2 which is somewhat odd to say the least, but power consumption is very impressive. It uses less power then the Opteron 285 even running 400 MHz faster. Using AMD's powernow! thing, the Opteron used less power then the Woodcrest, but that's b/c Woodcrest didn't have like half of the power saving options as it should have. It also seems to be a very realy stepping...
Click to expand...

True, but the insane gap between Conroe and AM2 based AMD processors most of the time was shown in SuperPI benchmarks and some ppl suggested that real world scenarios were not expected to be as extreme. Which is quite fortunate for AMD, as it would've been.... well, a massacre :D

I've been following core a lot, recently, and ever since the first benchmarks surfaced at IDF I was planning to get some woodcrest workstations for our studio. Now, IF these benchmarks are close to what is to be expected after the official launch, it will boil down to the better bang for buck ratio for me personally: If the AMD platform will be significantly cheaper due to memory etc. I think we will go for an Opteron based solution - I'll wait it out, I guess.

What really has to be considered IMHO, is that there is a high probability we can expect Cloverton to fit into the same boards, and that would definitely provide for some nice upgrade path by next year. Man, i cannot wait, bring it on Intel.... for christ's sake :p
 
Well if I remember correctly the platform calls for ddr2 667 fb dimms. and they were running ddr 2 533 fb dimms. I would guess that is some of the performance difference. Other than that meh, I have no idea.
 
pxc said:
I saw the link on /. this morning.

The 2 socket Woodcrest system tested was a little disappointing. It's still faster from the tech report's benchmarks of course, and the luckily the 5160 is $100 cheaper, runs cooler and is faster than the 285.

Maybe Poncho can give a little insight to the problem(s).
Click to expand...
I was thinking the same thing about the dissapointing results, but i wanted to be optmistic being that for the first time, Intel beat AMD in 2P systems.
 
Woodcrest had a sizable lead on a lot of the tests. I didn't think the results were disappointing. *shrug*
 
tys90 said:
Woodcrest had a sizable lead on a lot of the tests. I didn't think the results were disappointing. *shrug*
Click to expand...

QFT - nothing disappointing if you ask me. Just look at the rendering benchmarks, which shows a quite substantial peformance lead IMHO.
 
when you consider the fact that the Woodcrest had a 400 MHz clock speed lead over AMD, and the fact that it's new vs. old, it's dissapointing. But it was a very early CPU, so some of the bugs might have been fixed with the new stepping and performance might have been increased. Being that Poncho said woodcrest was a beast and he said the same thing about Conroe, performance has to be higher.
 
I guess it is a bit disappointing considering the MHz lead. Price comparision would be a better judge IMO. Performance/Price is better than AMDs for now, although the price of FB Dimms might make them about even (someone comment about that?). I can't say I follow the server market very well.
 
Duke3d87 said:
I was thinking the same thing about the dissapointing results, but i wanted to be optmistic being that for the first time, Intel beat AMD in 2P systems.
Click to expand...

First time lately, not first time ever. Intel had dual-CPU boards far before AMD entered the server market, and the Athlon MPs were not very impressive other than from a price point of view.

I figured it was memory bandwidth holding it back as well, so many server apps depend so much on bandwidth that maybe HT is giving the Opterons more breathing room, I'm sure of the Core Xeons had it, we would have seen much larger gaps.
 
NulloModo said:
I figured it was memory bandwidth holding it back as well, so many server apps depend so much on bandwidth
Click to expand...

You're onto something there. ;)
 
NulloModo said:
First time lately, not first time ever. Intel had dual-CPU boards far before AMD entered the server market, and the Athlon MPs were not very impressive other than from a price point of view.

I figured it was memory bandwidth holding it back as well, so many server apps depend so much on bandwidth that maybe HT is giving the Opterons more breathing room, I'm sure of the Core Xeons had it, we would have seen much larger gaps.
Click to expand...

It should be better then the Core series. Woodcrest has dual independant front side busses sharing up to 21 GB/sec of FBDIMM which exceeds the Core series. Plus the fact that you have 4 MB of L2 cache should help performance if Woodcrest is bottlenecked. Still, HT would boost performance quite a bit imo.
 
Duke3d87 said:
Woodcrest has dual independant front side busses sharing up to 21 GB/sec of FBDIMM which exceeds the Core series.
Click to expand...

Bensley has dual FSBs, not the processor. That's probably what you mean, but people will take it to mean that there are 4 FSBs. :D
 
tys90 said:
Woodcrest had a sizable lead on a lot of the tests. I didn't think the results were disappointing. *shrug*
Click to expand...
The results were with 3GHz Woodcrests vs 2.6GHz Opterons. Intel and AMD are both capable of releasing higher speeds. I did list the reasons why the Woodcrests were still good ("the 5160 is $100 cheaper, runs cooler and is faster than the 285").
 
Poncho said:
Bensley has dual FSBs, not the processor. That's probably what you mean, but people will take it to mean that there are 4 FSBs. :D
Click to expand...
yeah that's why i meant. I was looking at the diagram when i typed that, but yeah things happen..


About the HT support - it says yes b/c Dempsey supports HT...

So poncho, is the performance of the previewed woodcrest so "low" becuase it was an early model?
 
Duke3d87 said:
So poncho, is the performance of the previewed woodcrest so "low" becuase it was an early model?
Click to expand...


Not going to get into that discussion sorry. :D
 
Hmmm

http://www.computerhq.com/Supermicro_SUP_MB_X7DA8_RETAIL/X7DA8-O/hardware/partinfo-id-621098.html

$690 for a X7DA8 (5000X chipset)
$1700 for a pair of 5160s
$800 for 4GB of FB DIMMs
$big$ PSU
-------
= i've heard of people wasting more than that on a computer

The X7DA8 has 2 x16-sized PCI-E slots, one at x16 and the other at x4. That would still work for SLI/XF, right? I mean if someone wanted to really spend $5000 on a computer they could get something pretty fast without a huge chunk of the inflated price going to Alienware or others.
 
pxc said:
Hmmm

http://www.computerhq.com/Supermicro_SUP_MB_X7DA8_RETAIL/X7DA8-O/hardware/partinfo-id-621098.html

$690 for a X7DA8 (5000X chipset)
$1700 for a pair of 5160s
$800 for 4GB of FB DIMMs
$big$ PSU
-------
= i've heard of people wasting more than that on a computer

The X7DA8 has 2 x16-sized PCI-E slots, one at x16 and the other at x4. That would still work for SLI/XF, right? I mean if someone wanted to really spend $5000 on a computer they could get something pretty fast without a huge chunk of the inflated price going to Alienware or others.
Click to expand...

Well, at least we now know why Dell only talked about 4-way AMD systems.
 
Donnie27 said:
Well, at least we now know why Dell only talked about 4-way AMD systems.
Click to expand...
AMD is jumping on the FBDIMM bandwagon. They are letting Intel use the memory until they feel that it's cost effective enough. Several companies have announced intentions to mass produce FBDIMM so the prices should go down a bit. I hope Intel lowers the power consumption of the FBDIMM for overall improvements and potential frequency scaling.

And poncho, i'll take that as a more then likley yes.
 
Duke3d87 said:
And poncho, i'll take that as a more then likley yes.
Click to expand...


Don't take that to mean anything. Me not talking about it doesn't mean you are right wrong. Just means I'm not talking about it becuase that would be going a bit too far for my tastes. :D
 
Duke3d87 said:
AMD is jumping on the FBDIMM bandwagon. They are letting Intel use the memory until they feel that it's cost effective enough. Several companies have announced intentions to mass produce FBDIMM so the prices should go down a bit. I hope Intel lowers the power consumption of the FBDIMM for overall improvements and potential frequency scaling.

And poncho, i'll take that as a more then likley yes.
Click to expand...

You mean letting Intel and the rest Do the R&D while they take the cheapskate route. You know that's frowned appond by some in the Industry, right? Many companies doing that Letting other bring forth the tech) are seen as undependable. While folks on many message boards hate Intel, some industry folks fear having to depend on folks who DON'T invest a lot in the industry as a whole. I hope I made sense here.

As for FBDIMM, there are not Micron and Infineon backing stabbing the Tech as they did with RAMBUS. Of Course the Tech will see the light of day and do well.
 
jebo_4jc said:
Wow what happened to the huge IPC advantage.
Click to expand...

Or the need for an IMC LOL! This is just a wider Northbridge that can hander two FSBs, I wonder what Multiple Northbrides from Summit or etc.. would do?
 
Donnie27 said:
You mean letting Intel and the rest Do the R&D while they take the cheapskate route. You know that's frowned appond by some in the Industry, right? Many companies doing that Letting other bring forth the tech) are seen as undependable. While folks on many message boards hate Intel, some industry folks fear having to depend on folks who DON'T invest a lot in the industry as a whole. I hope I made sense here.

As for FBDIMM, there are not Micron and Infineon backing stabbing the Tech as they did with RAMBUS. Of Course the Tech will see the light of day and do well.
Click to expand...
wasn't samsung part of the group that pricefixed rambus? I'm pretty sure it was them too.

Donnie27 said:
Or the need for an IMC LOL! This is just a wider Northbridge that can hander two FSBs, I wonder what Multiple Northbrides from Summit or etc.. would do?
Click to expand...
It's not that. I still think there's something really wrong with those benchmarks. You take a CPU that has access to more memory and bandwith then its single CPU brother and you get a smaller performance jump. There's something wrong.
 
Donnie27 said:
Or the need for an IMC LOL! This is just a wider Northbridge that can hander two FSBs, I wonder what Multiple Northbrides from Summit or etc.. would do?
Click to expand...
I was actually talking about the IPC advantage Woodcrest was supposed to have over K8.
I wonder what happened to that?
 
jebo_4jc said:
I was actually talking about the IPC advantage Woodcrest was supposed to have over K8.
I wonder what happened to that?
Click to expand...
it's still there. There's more to performance then just IPC.If you look at the Intel released benchmarks and the other public benchmarks on Intel's site, you'll see a larger performance delta.
 
Donnie27 said:
You mean letting Intel and the rest Do the R&D while they take the cheapskate route. You know that's frowned appond by some in the Industry, right? Many companies doing that Letting other bring forth the tech) are seen as undependable. While folks on many message boards hate Intel, some industry folks fear having to depend on folks who DON'T invest a lot in the industry as a whole. I hope I made sense here.
Click to expand...

I agree with what you said. Those companies really don't bring anything good to an industry
 
I don't like the implication that AMD doesn't contribute to the Industry and just takes the cheap rout to copy Intel's tech.

Who introduced AMD64 and who coppied the instruction set?
Who set up the open HyperTransport standard that is used by dozens of companies world wide?

Just like Intel and any other smart company, AMD uses what it needs if it is available and invents it if it isn't.
 
jebo_4jc said:
I was actually talking about the IPC advantage Woodcrest was supposed to have over K8.
I wonder what happened to that?
Click to expand...

I didn't try to change what you said. I was talking about the so called bandwidth advantage Opteron has, NOT IPC of either :rolleyes: 12.6GB X 2 over 21GB. FSB/s Northbridge vs. IMC, sheesh!

Then to add to that, every app can't take advantage of IPC and that's been shown before. See Pentium-M for A64 during Games from two years ago for example?
 
Duke3d87 said:
It's not that. I still think there's something really wrong with those benchmarks. You take a CPU that has access to more memory and bandwith then its single CPU brother and you get a smaller performance jump. There's something wrong.
Click to expand...

I was thinking the same thing. You have a 2.66 Ghz Conroe destroying a FX60 OCed to 2.8 Ghz in some benchmarks, while in this case it's two 3.0 Ghz Woodcrests, which are essentially the same chip with a different memory interface, beating two 2.6 Ghz Opterons by less of a margin, even though an Opteron 2xx is just an A64 with a different name and dual Proc. capability.

When all of the other benchmarks are showing Core 2/Core Xenon chips a lot better off than this one, I'd tend to say Tech Report botched something in their review (and looking at the dual Opty system having double the RAM of the others, there is at least one clear sign they weren't that careful from the get-go).
 
You must log in or register to reply here.
Back
Top