AMD Athlon 64 FX-74 & Quad FX Platform Review

Paul_Jastrzebski said:
Thats actually a really interesting idea.. I don't have any server motherboards around but I will look into that. Even testing the new socket 1207 opterons might be interesting too...
Click to expand...
The AT review said this about putting opterons on the QuadFX mobo, I assume the opposite is true
"...for now just know that Quad FX only works with 1207-pin Athlon 64 FX CPUs (Opterons are prevented from working in the BIOS)."
I guess that AMD will make damn sure that no server motherboard maker releases a bios that supports the FX, and destroys the opteron market.

The AT review also said that CNQ isn't working on the board, which answers my earlier question about power consumption.
 
IanG said:
The AT review said this about putting opterons on the QuadFX mobo, I assume the opposite is true
"...for now just know that Quad FX only works with 1207-pin Athlon 64 FX CPUs (Opterons are prevented from working in the BIOS)."
I guess that AMD will make damn sure that no server motherboard maker releases a bios that supports the FX, and destroys the opteron market.

The AT review also said that CNQ isn't working on the board, which answers my earlier question about power consumption.
Click to expand...

So they are the same ;). Well, you can always try an older Socket F board. Not like AMD can magically retrofit old boards so they can't accept the new FX-7x series. I'm really anxious to see.
 
IanG said:
The AT review said this about putting opterons on the QuadFX mobo, I assume the opposite is true
"...for now just know that Quad FX only works with 1207-pin Athlon 64 FX CPUs (Opterons are prevented from working in the BIOS)."
I guess that AMD will make damn sure that no server motherboard maker releases a bios that supports the FX, and destroys the opteron market.

The AT review also said that CNQ isn't working on the board, which answers my earlier question about power consumption.
Click to expand...



Xbitlabs got CnQ to work and said this about power consumption. "Even in idle mode with enabled Cool’n’Quiet technology (without the monitor power consumption of course) this system consumed 192W of power, which is more than most contemporary systems would require for 100% workload."Xbitlabs
 
IanG said:
The AT review said this about putting opterons on the QuadFX mobo, I assume the opposite is true
"...for now just know that Quad FX only works with 1207-pin Athlon 64 FX CPUs (Opterons are prevented from working in the BIOS)."
I guess that AMD will make damn sure that no server motherboard maker releases a bios that supports the FX, and destroys the opteron market.

The AT review also said that CNQ isn't working on the board, which answers my earlier question about power consumption.
Click to expand...
well u could mod the bios.
 
I have to wonder if AMD was even aware of the overall power consumption of the system company wide. I had one of the FX project managers ask me last night, what I thought and this was my reply.

Power consumption and cost are going to be killers till you can leverage your quad core next year…..
Click to expand...

His reply:

Actual power for us is around 95w per proc and 140w for the competitors.

50w, is that such a big deal?

Also, you can get into quad core with us for 599, check out kentsfields Newegg pricing today!

You can build a QuadFX system for 200$ more then a kentsfield costs!!
Click to expand...

I sent him our power graphs that are in the article and asked him to call me to discuss and I never heard back from him at all. Given that none of these are for sale yet it hard to argue real costs associated with building them.
 
mzs_biteme said:
More on power usage:



http://www.legitreviews.com/article/425/15/

Extreme situations (QuadFX) call for extreme measures.... :eek:
Click to expand...

Yeah, looks like a nuclear power plant in your basement wouldn't be a bad idea for those considering QuadFX :eek:

I can't justify the positions taken by some that the power useage is acceptable or tolerable, particularly given these same people's beratement of the Pentium EE back in the day.
 
DryFire said:
I remember AMD passing out oven mitts with the prescott came out, I wonder if Intel will return the favor now.
Click to expand...

ROFL, oven mitts? Intel may start handing out fire extinguishers instead!
:D
 
Martyr said:
well u could mod the bios.
Click to expand...
I was under the impression it was physical limitation in that Opteron requires registered RAM, and that QuadFX requires non-registered RAM
 
I have to say that a site that prides itself on real world testing seems to be scrambling a bit to try and test the FX 4x4 and QX6700 chips. Could it be that outside of a render farm there is no practical use for these chips yet?

I do realize that the hardware must exist before there is software to take advantage of it but, I see no 'practical' advantage outside of an engineering exersize for these chips for at least 2 YEARS!

I guess I'll have to turn in my early adopter badge. To all the folks who just have to have one now I can only say "Bleed Baby, Bleed!"
 
DryFire said:
I remember AMD passing out oven mitts with the prescott came out, I wonder if Intel will return the favor now.
Click to expand...

LMAO, that would be mean! But hey payback is a bitch.
 
AMD's rep quote:

Actual power for us is around 95w per proc and 140w for the competitors.
50w, is that such a big deal?
Also, you can get into quad core with us for 599, check out kentsfields Newegg pricing today!
You can build a QuadFX system for 200$ more then a kentsfield costs!!

Abviously some poeple don't do math well: 2x95W CPU vs 1x140W CPU is not 50W difference...
So now they're comparing their LOWEST performing $599 set of QuadFX CPU's against top of the line Intel QX.... Anybody wanna notice here that at this very moment QuadFX is a paper launch, while the demand for QX is so high that it's always out of stock even at $1500... :eek: By the time you'll actually be able to get QuadFX, Intel will have released QX6600 and QX6400 CPU's... Aslo remember, Intel is set to cut their prices in Jan '07.
It's funny so see how AMD is coming out with some twisted logic to justifiy QaudFX's existance at this very place in time.. :confused: :cool:
 
chris.c said:
That's my point Kyle, when I got to the conclusion page I wondered to myself if we had read the same review.

What I saw was AMD getting bested by Intel in the great majority of cases, while consuming twice as much power. So I was surprised to read this:

"[We had doubts], but our real world testing today disproves my preconceptions entirely, and shows that in quite a few cases, the FX-74 is as fast, or even faster than the QX6700. We ran the numbers over and over again, and the FX-74 simply has the horsepower to compete with the QX6700 in the most CPU intensive applications."

In addition, your previous "bottom line" was no help at all. The amendment has made it clear, so thanks for that.

What it comes down to is, you review hardware for a living, I don't. So when I read the article and came to an entirely different conclusion than you did, I wanted to see how you came to said conclusion.
Click to expand...

As noted, that has been fixed for hours. Thanks for bringing up your points.
 
You know, reading this I couldn't help but think that maybe it's time for a new slogan around here. Even with the "appended" conclusion, which basically does little other than trying to pretend AMD is going to be the better future platform (i.e. if you don't plan on upgrading), there is the typical rampant AMD bias. Too close to declare a winner? Of course! Kyle is much more forthcoming in the forums, where he jokes about the power draw and other problems, but why the hell is the article conclusion and other aspects so damnably flattering!? Which brings me to my new slogan:

HardOCP: Trying to force the "[H]" into "Advanced Micro Devices!" :D

As I was saying, the "amended" conclusion suggests AMD will provide the much better platform in the future. Obviously, Kyle has a much better crystal ball than the rest of us, because given the massive overclocks people are achieving with Kentsfield processors I seriously doubt that AMD's only competition is going to be the QX6700. All Intel needs to do is release a Xeon chip set that can use unbuffered (or even registered!) memory instead of the lousy FB-DIMMs, and they would have an octal core "desktop" platform available today. The conclusion should have been more like the following:

"AMD would like us to believe that later this year, Quad FX will be the best computing platform available. They could very well be right, but even so we would recommend waiting until that platform is actually available in its final incarnation rather than wasting your money today on a platform that is -- at best -- roughly equal to Intel's Kentsfield platform. Just as no one in their right mind would have recommended purchasing one of the Pentium Extreme Edition plus HyperThreading processors over the cooler running and faster Athlon 64 X2/FX, we are not going to recommend an overall slower, more power-hungry Quad FX system over Intel's Kentsfield. Unless of course we need a new truck, in which case AMD and Quad FX rock!"

Not your momma's PC indeed... because up until the launch of Core 2 Duo, AMD was providing faster, cooler running, more efficient processors. But, of course, the testing "disproves [Kyle's] preconceptions entirely." Your conclusion disagrees with itself, Kyle. You can't edit one paragraph and fix a completely broken conclusion.

PS -- Posted from a socket 939 Opteron 165 @ 2.4GHz system. (Would be 2.8 GHz if it weren't for this lousy ATI Xpress 3200 chipset!)
 
Magnus said:
I have to say that a site that prides itself on real world testing seems to be scrambling a bit to try and test the FX 4x4 and QX6700 chips. Could it be that outside of a render farm there is no practical use for these chips yet?

I do realize that the hardware must exist before there is software to take advantage of it but, I see no 'practical' advantage outside of an engineering exersize for these chips for at least 2 YEARS!

I guess I'll have to turn in my early adopter badge. To all the folks who just have to have one now I can only say "Bleed Baby, Bleed!"
Click to expand...


I would rather scramble and stay in the real world, rather than vomit canned benchmarks that mean little if anything. :)

Direct to your point, there are few people that need a quad-core processor. I think we have been very clear about that over and over.
 
frumious1 said:
HardOCP: Trying to force the "[H]" into "Advanced Micro Devices!" :D
Click to expand...


:eek: :eek:

I thought the article was AMD biased after reading the C2D review again looking for reasons at to why the original conclusion was stated. Probably not that biased but overall there is no reason to believe this platform is a real competitor to the Kentsfield platform at this time. While people are bringing up the point of a 3GHz CPU losing to a 2.66GHz CPU, is it not really 6GHz losing out to 2.66GHz. ;) Oh well, I expect we will be saying the same things come next summer only in reverse.
 
frumious1 said:
.... Which brings me to my new slogan:

HardOCP: Trying to force the "[H]" into "Advanced Micro Devices!" :D

....
Click to expand...

:eek: :D :eek:

Damn you!!!! You made me spew my Coke all of over the keyboard...
 
mzs_biteme said:
Abviously some poeple don't do math well: 2x95W CPU vs 1x140W CPU is not 50W difference...
Click to expand...

When did 190-140 stop being 50?? WHY WASN'T I NOTIFIED?! People's lives depend on this!!
 
You can build a QuadFX system for 200$ more then a kentsfield costs!!.
Click to expand...

And even if that $200 price difference weren't so glaring, the yearly power cost trumps that even more.

Tom's had the 4x4 at 0.24kwh per 30 minutes when running 3DMark06. Double and round that to 0.50kwh per 60 minutes. If you pay the US average $0.10 per kwh, run your computer 24/7 over a year and that's $438/year in power costs. OK even if you run 8 hours per day, that's still $146/year.

Conversely I pay around $0.12 per kwh and with a meter, my hot Prescott system pulls just shy of 300W from the wall WITH LCD and UPS when running 3DMark06, and costs me roughly $0.04 per 75 minutes. 8 hours a day and that's $93.44 a year in power.
 
FlintMurdock said:
Which benchmarks were canned? All the benchmarks I saw had pretty much the same info. Only the opinions and conclusions seemed to differ.
Click to expand...

I am just telling you what we do not want to do on our own site. I am making no reference to anyone else's work.
 
phobos512 said:
When did 190-140 stop being 50?? WHY WASN'T I NOTIFIED?! People's lives depend on this!!
Click to expand...
That's what on the surface... people fail to notice that 2x95W CPU needs 2x the CPU Socket = 2x the power circuits, demending 2x MOSFETs, filters, etc, leaking 2x as much power.... :eek: Would be nice if you could just cut that middle crap out... ;) :D
 
mzs_biteme said:
That's what on the surface... people fail to notice that 2x95W CPU needs 2x the CPU Socket = 2x the power circuits, demending 2x MOSFETs, filters, etc, leaking 2x as much power.... :eek: Would be nice if you could just cut that middle crap out... ;) :D
Click to expand...

No I wasn't neglecting those things BUT the guy was specifically talking about the difference in processor power, not system power.

From the caliber response that Kyle received from the guy and posted above I'd imagine the guy was not one of their engineers (or more correctly, I PRAY THAT HE IS NOT). He'd be the marketing guy or something - that's why he doesn't care. It's not that he doesn't care, he probably just doesn't have a clue.
 
mzs_biteme said:
Abviously some poeple don't do math well: 2x95W CPU vs 1x140W CPU is not 50W difference...
Click to expand...

Let's have a little maths lesson.

2 x 95W = 190 W
1 x 140W = 140 W
----------------------------
Difference = 50 W


Lol.
 
theelectic said:
And even if that $200 price difference weren't so glaring, the yearly power cost trumps that even more.

Tom's had the 4x4 at 0.24kwh per 30 minutes when running 3DMark06. Double and round that to 0.50kwh per 60 minutes. If you pay the US average $0.10 per kwh, run your computer 24/7 over a year and that's $438/year in power costs. OK even if you run 8 hours per day, that's still $146/year.

Conversely I pay around $0.12 per kwh and with a meter, my hot Prescott system pulls just shy of 300W from the wall WITH LCD and UPS when running 3DMark06, and costs me roughly $0.04 per 75 minutes. 8 hours a day and that's $93.44 a year in power.
Click to expand...

Where they'll really get you is when the AMD system throws you over your seasonal baseline and you start paying double or more per energy unit used then if you'd been under baseline.
 
phobos512 said:
No I wasn't neglecting those things BUT the guy was specifically talking about the difference in processor power, not system power.

From the caliber response that Kyle received from the guy and posted above I'd imagine the guy was not one of their engineers (or more correctly, I PRAY THAT HE IS NOT). He'd be the marketing guy or something - that's why he doesn't care. It's not that he doesn't care, he probably just doesn't have a clue.
Click to expand...


Seems as if AMD is nothing but Marketing BS when it comes to QuadFX... And where all buying into it...right.... :rolleyes: ;) :D
 
"Actual power for us is around 95w per proc and 140w for the competitors.
50w, is that such a big deal? "


I love how he plays this off as its not a big deal, but if it was the other way around he would be all over it.
 
D4hPr0 said:
"Actual power for us is around 95w per proc and 140w for the competitors.
50w, is that such a big deal? "


I love how he plays this off as its not a big deal, but if it was the other way around he would be all over it.
Click to expand...

Finally someone else see's this also.
 
D4hPr0 said:
"Actual power for us is around 95w per proc and 140w for the competitors.
50w, is that such a big deal? "


I love how he plays this off as its not a big deal, but if it was the other way around he would be all over it.
Click to expand...

No kidding... He'd probably start: "Mittens for Intel #2" project.... :rolleyes: Talk about a cheap-shot...
 
mzs_biteme said:
No kidding... He'd probably start: "Mittens for Intel #2" project.... :rolleyes: Talk about a cheap-shot...
Click to expand...

Now that this thread has been overthrown by the zealots and !!!!!!s, I think I'll take stage left.
 
Clearly, in WinXP, QuadFx is no match for kentsfield. You can argue about how unclear or how clear the review was, but that's your own problem. The benchmarks clearly state Intel is ahead here. I don't mean the architecture or the elegance or w/e, just the numbers.
However, due to the lack of real NUMA/memory/all that other technical nonsense not fully supported by WinXP, how come no one is running Vista benchmarks (or at least a preview on RTM or something, feeeeeeed meeeee!)?
Let's see what AMD can really do when the limitation is not some preliminary BIOS or an OS that wasn't designed to handle this type of hardware. After we see what happens once the architecture is fully supported by the OS, then let's have pointless arguing over whose Hummer is better than whose Kentsfield (think about it). Not to mention that AMD explicitly stated that the BIOS is not fully functional yet and has a glitch concerning performance "In particular, AMD told us that the current 0117 BIOS has a bug which affects synthetic memory benchmarks. The NUMA system is also foreshortened by problems with the BIOS's Static Resource Affinity Tables."

Back to the whining!
/CHARGE!
 
PunjabiPlaya said:
Clearly, in WinXP, QuadFx is no match for kentsfield. You can argue about how unclear or how clear the review was, but that's your own problem. The benchmarks clearly state Intel is ahead here. I don't mean the architecture or the elegance or w/e, just the numbers.
However, due to the lack of real NUMA/memory/all that other technical nonsense not fully supported by WinXP, how come no one is running Vista benchmarks (or at least a preview on RTM or something, feeeeeeed meeeee!)?
Let's see what AMD can really do when the limitation is not some preliminary BIOS or an OS that wasn't designed to handle this type of hardware. After we see what happens once the architecture is fully supported by the OS, then let's have pointless arguing over whose Hummer is better than whose Kentsfield (think about it). Not to mention that AMD explicitly stated that the BIOS is not fully functional yet and has a glitch concerning performance "In particular, AMD told us that the current 0117 BIOS has a bug which affects synthetic memory benchmarks. The NUMA system is also foreshortened by problems with the BIOS's Static Resource Affinity Tables."

Back to the whining!
/CHARGE!
Click to expand...

because everyone wants to use the 8800gtx, but nvidia hasnt seen fit to release vista drivers for it yet
 
I can't wait when Barcelona comes out and I revive this thread for kicks.

Bitch and moan all you like about these 3 year old architecture cores (that are still neck and neck with Intel's latest) that suck up double the power. Although how many here OVERCLOCK their processors and setup contraptions that would make NASA blush, yet still bitch about power and heat :rolleyes:

This is [H] last time I checked. I never seen so many people get into a tissy over a temporary processor solution that sucks up a little more juice compared to Intel's THIRD RESPONSE to K8 that came out only two months ago :rolleyes: . I got the money to blow on this setup you think I give a rats ass about power consumption, especially since I only need it for 5 months until Barcelona.

Bottom line, AMD with it's old architecture and it's incredible HT throughputs still can keep competitve with Intel's latest. So what are going to be the excuses when K8L comes out (you know the answer to CD2 and Intel's Quad)? ...I am going to love to hear the responses later on when Intel not only gets spanked with a straight up 4 core fight, but when they have to deal with an enthusiast 8 core system.

For the record I own and use a double dual core Opty 275 system and this Cowboy loves this new platform. Only wish they had one that was Crossfire capable.

Today was a day for showing off a new mobo platform, not "new" processors :rolleyes:
 
duby229 said:
Now that this thread has been overthrown by the zealots and !!!!!!s, I think I'll take stage left.
Click to expand...

Well this is are "revenege" for all the perfectly good threads in the Intel forum that got trashed by AMD f**boys back in the Pentium 4/D vs Athlon 64 days...... :rolleyes: Now you know exactly how it feels!!!
 
Lazy_Moron said:
Well this is are "revenege" for all the perfectly good threads in the Intel forum that got trashed by AMD f**boys back in the Pentium 4/D vs Athlon 64 days...... :rolleyes: Now you know exactly how it feels!!!
Click to expand...
Such weak justification..... But mommy! Johnny did it first! waaaaaaaaaaaa. Don't even pretend the crap that happens between the forums is one sided.


Kyle:

Great article, sorry you have to defend everything you ever write around here. I appreciate it though.



Hopefully this is just a stopgap until AMD can get something better out. The power consumption on this system is well beyond acceptable levels for me. As of today, Intel still holds the better solutions in my opinion.

My AM2 4200+ on a Crosshair w/ SLI is fitting my needs well for the time being, hopefully I can get 6 to 8 more months out of it before I feel the overwhelming need to boost it again.
 
|CR|Constantine said:
....
Today was a day for showing off a new mobo platform, not "new" processors :rolleyes:
Click to expand...

Oh yeah, sure it's not about the CPU's... Sorry... Stupid me... It's all about the mobo platform that ALL mobo manufacturers are so excited about... Riiiight... :rolleyes:
 
I think every site that has reviewed the quad FX has totally cheated AMD.

NUMA is the main benefit to having two separate CPUS with their own dual channel controllers. At least run some Linux benchmarks that have NUMA support to show real performance of the system.

Technicly the quad FX blows the intel solutions outa the water becuase you can run 4 graphics cards, at a price and power of course.

When vista is out with NUMA support we'll see how well these 90mm chips can really shine.
 
Unoid said:
I think every site that has reviewed the quad FX has totally cheated AMD.

NUMA is the main benefit to having two separate CPUS with their own dual channel controllers. At least run some Linux benchmarks that have NUMA support to show real performance of the system.

Technicly the quad FX blows the intel solutions outa the water becuase you can run 4 graphics cards, at a price and power of course.

When vista is out with NUMA support we'll see how well these 90mm chips can really shine.
Click to expand...
This platform is marketed to gamers. Gamers do not use Linux. So not using Linux is fully justified.

And as for Vista, it's a little hard to test hardware on an operating system that's not available yet. ;)
 
Unoid said:
I think every site that has reviewed the quad FX has totally cheated AMD.

NUMA is the main benefit to having two separate CPUS with their own dual channel controllers. At least run some Linux benchmarks that have NUMA support to show real performance of the system.

Technicly the quad FX blows the intel solutions outa the water becuase you can run 4 graphics cards, at a price and power of course.

When vista is out with NUMA support we'll see how well these 90mm chips can really shine.
Click to expand...
we can already run quad cards. the advantages of doing so are few.
 
mzs_biteme said:
Oh yeah, sure it's not about the CPU's... Sorry... Stupid me... It's all about the mobo platform that ALL mobo manufacturers are so excited about... Riiiight... :rolleyes:
Click to expand...

Mobo manufactures excited enough to get pissed off at AMD for launching with only Asus so yea I would say they are excited.

What the hell is so special about a K8 clocked to 3.0GHz :rolleyes: The real story here is the new mobo platform and the ability to run two of the K8L's in the future.

But of course you would rather talk about how K8 does against Intel's third and fourth response to it. Even though it's neck and neck in the benches and according to Kyle the exact same in real performance.

I like the Quad FX system and I happy AMD ONCE AGAIN is pushing boundries.



How is that FSB and massive cache coming along?...
 
You must log in or register to reply here.
Back
Top