They have, just wait.Quote:
Originally Posted by Ubermann
Printable View
They have, just wait.Quote:
Originally Posted by Ubermann
Quote:
Originally Posted by biohead
.....yea just another year or so.....:)
funny how a FX-60 outperformed the 62 there. Oh well, probably just a fluke.Quote:
Originally Posted by ahmad
The FX-62 system runs at 2T, and I think in games DDR1 has the edge unless the DDR2 is at an xtreme clock.Quote:
Originally Posted by Cybercat
These seemed to be the interesting results:
Quake 4 - 1600x1200 - 4x AA 16x AF (higher is better) 115.8 113 111.7 107.23 108
Splinter Cell: CT - 1600x1200 - 4x AA 8x AF (higher is better) 63.87 64.01 63.31 62.33 62.05
As long as we remain GPU limited the power of the CPU is mostly irrelevant.
Well, we knew that -- no matter what the platform.Quote:
Originally Posted by Bar81
What's interesting to me is this:
ScienceMark 2.0 memory latency (ns, lower is better) 45.37(E6700) 43.31(E6600) 44.41(FX62) 47.83(FX60) 70.11(965)
Is there something screwy with ScienceMark's way of benching memory latency? Or can Conroe's FSB approach really be that efficient?
Nice little review
http://techreport.com/etc/2006q2/woodcrest/index.x?pg=1
It's funny how all these comparison's between the "new" Intel and AMD comes out right after AMD releases it. Intel has a very very very smart marketing team.
http://techreport.com/etc/2006q2/woodcrest/cpuz-cpu.gif
http://techreport.com/etc/2006q2/woo...nsley-guts.jpg
http://techreport.com/etc/2006q2/woo...ine-render.gif
http://techreport.com/etc/2006q2/woodcrest/cine-c4d.gif
http://techreport.com/etc/2006q2/woodcrest/povray.gif
http://techreport.com/etc/2006q2/woo...x-scanline.gif
http://techreport.com/etc/2006q2/woodcrest/wme-ap.gif
http://techreport.com/etc/2006q2/woodcrest/xmpeg.gif
http://techreport.com/etc/2006q2/woo...ame-cbr-ms.gif
http://techreport.com/etc/2006q2/woo...-cbr-intel.gif
http://techreport.com/etc/2006q2/woo...ame-vbr-ms.gif
http://techreport.com/etc/2006q2/woo...-vbr-intel.gif
http://techreport.com/etc/2006q2/woodcrest/sphinx.gif
http://techreport.com/etc/2006q2/woodcrest/picc.gif
http://techreport.com/etc/2006q2/woo...picc-indiv.gif
http://techreport.com/etc/2006q2/woo...dra-mm-int.gif
http://techreport.com/etc/2006q2/woo...ndra-mm-fp.gif
http://techreport.com/etc/2006q2/woo...uz-latency.gif
http://techreport.com/etc/2006q2/woo...power-idle.gif
http://techreport.com/etc/2006q2/woo...power-load.gif
Quote:
It's taken some time, but Intel finally has a strong answer to the Opteron—or nearly so. The Bensley platform provides a much-improved infrastructure for low-end and mid-range servers, with vastly more bandwidth to each CPU socket—and to main memory—than previous Xeon chipsets. When they arrive, Woodcrest-based Xeons look like they will allow Intel to recapture the performance lead in server and workstations-class CPUs. We obviously haven't tested server-class applications yet, but the performance lead we've seen from Woodcrest in nearly every test we have run bodes very well indeed. So do Woodcrest's power consumption numbers, which are a revelation compared to what we've seen from Netburst processors in recent years.
Of course, when Woodcrest-derived Xeons do arrive, they may be contending with the upcoming revision-F Opterons with DDR2 memory, lower system power consumption, and perhaps higher clock speeds, as well. Will these changes be enough to keep AMD competitive against Woodcrest? That's hard to say for sure, but it certainly doesn't look like life will be easy for AMD.
good to see the new woodcrest @ 3.0 beating the old amd opteron @ 2.6, oh, not by much it seems
Is woodcrest the new 4 core CPU? Or am I getting it confused with another CPU upcoming from Intel?
I would read the article but the link isn't loading for me.
Yeah, normally companies stay hush hush about their upcoming products, you hear a little snippet here and there, but never like we've seen with Merom and Conroe.....
Intel knew they had no choice but to put conroe out everywhere. The chips in peoples hands were intentionally leaked, the benchmarks were with permission from intel. Intel had no choice, they HAD to stop the mass consumption of AMD chips over the past 6 months, no if's, and's, or but's about it.
They did what they had to, and it looks like it's going to pay off in a big way. Atleast amongst the desktop enthusiasts. Servers, where the REAL money is, will still remain all amd's game.
Finally, to the aMp, the reason for that anomoly in memory latency was the fact that the intel systems were using different ram than the FX-62(3-2-2-8 2t@667 vs 4-3-3-8 2t@ 800mhz for the amd). So yes, i'd expect lower memory latency out of the Intel,considering it's using low latency ram.
100+?Quote:
Hi everyone!
was wondering about the prices of those infinity series mobos for Conroe - any ideas how much they will cost?
Those, whom do not likes memory timings of FX-62 in this article, can go and read the appropriate article, compare it with results of Conroe and reveal that better timings for FX-62 do not change picture significantly.
Quote:
Originally Posted by CedricFP
WOODCREST is single die, dual core XEON from intel, based on Conroe and designed for dual socket servers, so results are form 4 core 2 processor, pre relelase server:D
For those of you who prefer a graph layout, like me, I graphed the results. See below:
http://img333.imageshack.us/img333/4...roevam22iw.jpg
Why didn't you scale it from 0 on the last two? It doesn't need to be exagerated if they're that close. The game benchmarks are mixed, but it looks like A64s should be fine with gpu intensive applications. I'll wait for 2007 and quadcore to see what cards AMD and Intel really have.
But both the Presler and Conroe ran 3-2-2-8 @ 667MHz. Same memory speed and timings, but Conroe's latency is something like 35 percent lower. That's huge.Quote:
Originally Posted by DilTech
Is there some architectural difference that accounts for the difference? I wouldn't have thought that sort of improvement could be achieved without moving to an on-die mem controller.
Not trying to be a jerk. It's entirely possible that I just have no idea what I'm talking about. I'd just like to know why. :toast:
Processor speed makes a difference as well when it comes to memory bandwidth/latency, this has been shown more than a few times.
Do not also forgot, conroe has a MUCH lower latency cache than presler, which would further explain the difference.
Now to find the OC friendly server mobo :D
There gonna be 1.6, 1.8 GHz Woodcrests with 4MB L2 and 1066 FSB :slobber:
Maybe LV Woodcrests too?Quote:
Originally Posted by Cooper
Wonder if anyone has heard of Asus doing another xeon workstation board.
Didn't even notice that. It has been updated.Quote:
Originally Posted by r3w4
I hope this means they will start to make new games more CPU dependant... use all this new power
cool, that's pretty fast
Thx to Xs News Team member Pinnacle for the server roadmap:
http://dailytech.com/article.aspx?newsid=833
All we need now is the Desktop roadmap.
If anyone can find updated amd or intel roadmaps extending into 2007 pm me. Thread closed until further notice.
Perkam