TomsHardware: Core i7-3970X Extreme Review: Can It Stomp An Eight-Core Xeon?

[Idontcare]

Originally Posted by octavian33

We all know that Intel makes the faster processor. We also all know that intel is way more expensive. So my simple question is this, why doesnt amd just make a board that will take 2 FX-8350's. I mean is there a reason that 2 processors running in unison wouldnt at least or exceed intels 3970x in multithreaded applications, and at a cheaper price? I am a total newb at multi processor set ups but it seems logical to me.

They've been down this road before, both companies, of crafting 2-socket mobos for the enthusiast segment and the bottom line is that they just don't sell very well.

AMD did it and it was called the "quad-father" (two dual-core CPUs) at the time, it performed ok but it pushed the power usage numbers to ludicrous levels and thus became a pariah for that reason.

I suspect an enthusiast-level dual socket FX-8350 platform would be similarly castigated by the enthusiast community, sadly, and that concern alone probably torpedoes the idea at the drawing board.

For the HPC and workstation markets, where performance truly matters in the economic sense, overclocking is out anyways so those platforms are going to be based on Xeons and Opterons which already have their own dual-socket platforms.

It is just the minor segment of at-home OC'ers who are into this stuff and it would appear that for most of us when we say we are "into it" that really mean we "like to read about it", which translates into very few sales. If you know your platform volume is going to be low then it becomes harder and harder to justify the expense of crafting niche products like enthusiast-class dual-socket mobos (skulltrail for example).

[lamedude]

Originally Posted by bunnyfubbles

there are 8 core Xeons on 32nm, the problem is that they're completely locked down outside of the typical maybe couple extra MHz on BCLK overclocking like we see with Sandy and Ivy 1155 non K CPUs, and thus a heavily overclocked 6 core is actually going to be faster than the fastest 8 core Xeon and completely not worth it for an enthusiast who is willing to blow past stock TDP with proper cooling.

SNB-E can do 125,166, or 250MHZ BCLCK.
31-38*125=3.8-4.7GHZ

[Makaveli]

Originally Posted by octavian33

We all know that Intel makes the faster processor. We also all know that intel is way more expensive. So my simple question is this, why doesnt amd just make a board that will take 2 FX-8350's. I mean is there a reason that 2 processors running in unison wouldnt at least or exceed intels 3970x in multithreaded applications, and at a cheaper price? I am a total newb at multi processor set ups but it seems logical to me.

Do FX chips even have the correct hypertransport setup to support Dual socket workstation board?

That is usually the difference with the opteron and xeons!

[Nemesis 1]

Originally Posted by inf64

50% better performance? How about some facts from the article? Time to completion of whole THG test suite(ST and MT workloads) for 3970x is 91.5 minute. FX8350 took 116 minutes . From this we have: 91.5/116=0.788 or ~21% faster than 8350. 3770K is similarly 12% faster than 8350 when you take its total runtime in the equation.

YA i was a bit high there, I gave type of numbers you give befor the release of every AMD processor.

If I took a 3770K and overclocked it to the same power usage as the amd . The intel cpu would be running around 4.4ghz when you equalize power out of the socket .. The Best metric is power usage . Time of Work is user choice but intel would have a huge advantage here. Now you could argue were talking stock. But than I would point to Amd sales numbers and profits. Ya the 8 core amd chip is fast but its a power hog. For the same power usage intel users have way better performance In hardware forums that is. The people speak with their money . Thats not going so well for Amd

[njdevilsfan87]

Originally Posted by lamedude

SNB-E can do 125,166, or 250MHZ BCLCK.
31-38*125=3.8-4.7GHZ

That is not something you can do on SNB-EP unfortunately.

[inf64]

Originally Posted by Nemesis 1

The people speak with their money . Thats not going so well for Amd

Funny that 8350 has been constantly sold out at newegg. So much for not going so well for AMD. I'm willing to say that AMD will exceed their sales expectation with Vishera line,far outselling FX8/6/4 series this and next year.

[AdamK47]

Originally Posted by lamedude

SNB-E can do 125,166, or 250MHZ BCLCK.
31-38*125=3.8-4.7GHZ

Unfortunately, base clocks of 125, 166, and 250 do not work on the Xeons. I'd have one if they did.

[AtenRa]

Originally Posted by Idontcare

AMD did it and it was called the "quad-father" (two dual-core CPUs)

The Quad Father was the first true x86 quad core CPU, the Phenom I (i have a T-shirt from Cebit 2008). The dual socket was in the Athlon FX era.

Edit: Nope i was wrong, it was the dual Socket

[Makaveli]

Originally Posted by inf64

Funny that 8350 has been constantly sold out at newegg.

Sales alone doesn't guarantee anything.

If you price stuff cheap enough people will buy it.

I'm not saying it isn't a good cpu for the price you pay.

You cannot look at sales numbers alone.

[inf64]

Well its price is positioned in the market "sweet spot" ** It's not overpriced nor under priced And debating whether AMD gets any money is absurd** I remember someone once did a calculation on production costs at Ace's HW forum and it turned out AMD payed from 45-65$ per good die** I see no reason why this would have changed especially now that AMD has zero investments in process nodes and only pays for good dies now**

When you sell that chip for ~150-200 bucks it certainly is not losing money** And then you sell same die in other form(server) for 5x more,you do the math whether they lose money or earn money** Margins are low and that is true ,but that is because AMD sells a lot of cheap CPUs(cheap dual and even quad core APUs for OEM machines) at bargain prices and they make a large part of their shipments**

PS Forum is acting up, we have a lot of ** in our posts**

[Edrick]

Originally Posted by inf64

When you sell that chip for ~150-200 bucks it certainly is not losing money

You are not taking into account all the R&D, packaging, overhead costs that go into making the product** Yes AMD is making some money, but not even close to where they want to be**

[Ben90]

I'm amazed they still see a market for 130W+ CPUs in the desktop space** CMON 200W Haswell!

[mrmt]

Originally Posted by inf64

Well its price is positioned in the market "sweet spot"** It's not overpriced nor under priced And debating whether AMD gets any money is absurd I remember someone once did a calculation on production costs at Ace's HW forum and it turned out AMD payed from 45-65$ per good die** I see no reason why this would have changed especially now that AMD has zero investments in process nodes and only pays for good dies now

(******)

When you sell that chip for ~150-200 bucks it certainly is not losing money**

First AMD is certainly not happy with GLF pricing and execution otherwise they wouldn't have shelled out 600 million to get out of the contract with them** That means bad yields or very high costs, to the point that AMD sees more than 600 million in gains from breaking up with GLF**

Second, your numbers for costs aren't too off mark but your revenues estimates are** Assumming that AMD sold 11 million units at the confirmed 38% gross margins, this means $54 COGS per chip and ASP of $87 per chip, servers included**

AMD certainly isn't losing money on the FX 8350 but you can be pretty sure that AMD cannot sell them to retailers and distributors at $150-$200, it's way less, plus I wouldn't be so sure on them making much money, if any, on the FX 4xx0 (which shares the same 8xx0 die) or the single module Trinity, which shares the same 246mm^2 5800k die**

[skipsneeky2]

Intel has popped another prescott out of the oven!

150wtdp and not very much quicker then its predecessor let alone a $600 3930k@#@#@#@#@#@#@#@#how is this thing even appealing with a whopping 150w tdp?

I sense a bulldozer @#@#@#ment in@#@#@#ing as most are gonna exclaim those are the prescott chips!

[Makaveli]

Originally Posted by skipsneeky2

Intel has popped another prescott out of the oven!

150wtdp and not very much quicker then its predecessor let alone a $600 3930k@#@#@#@#@#@#@#@#how is this thing even appealing with a whopping 150w tdp?

I sense a bulldozer @#@#@#ment in@#@#@#ing as most are gonna exclaim those are the prescott chips!

How is it another prescott when its the highest performing Desktop processor they have to offer?

Prescott never held that title@#@# And if we assume your argument is about heat prescott never offered the performance for its heat generated, you cannot say the same about SB-E@#@#

Extreme edition cpu's are ment for those that want the best of the best and can afford it@#@#

Its obviously not in the market for alot of us yet you don't see me going to the lambo forum and @#@#@#plaining about the price

[skipsneeky2]

Originally Posted by Makaveli

How is it another prescott when its the highest performing Desktop processor they have to offer?

Prescott never held that title@#@# And if we assume your argument is about heat prescott never offered the performance for its heat generated, you cannot say the same about SB-E@#@#

Extreme edition cpu's are ment for those that want the best of the best and can afford it@#@#

Its obviously not in the market for alot of us yet you don't see me going to the lambo forum and @#@#@#plaining about the price

Will reply with 2 words@#@#@#@#@#@#@#@#@#@#

150w tdp,price well the smart ones are buying a 3930k,like you were the smart one buying the 970 over the 980x@#@#

Twice the tdp of a 3570k@#@#@#@#@#@#@#@#a measily 200mhz jump over the 3960x@#@#

[ShintaiDK]

Atleast they dont have try and fake it by calling it ACP.

[Makaveli]

Originally Posted by skipsneeky2

Will reply with 2 words@#@#@#@#@#@#@#@#@#@#

150w tdp,price well the smart ones are buying a 3930k,like you were the smart one buying the 970 over the 980x@#@#

Twice the tdp of a 3570k@#@#@#@#@#@#@#@#a measily 200mhz jump over the 3960x@#@#

hehe I can't even see the reply with all these random characters in the post@#@#

someone please fix the forum@#@#

But I think were on the same page@#@#

[blckgrffn]

Originally Posted by Makaveli

hehe I can't even see the reply with all these random characters in the post@#@#

someone please fix the forum@#@#

But I think were on the same page@#@#

I thought it was me! Holy crap is it goofed to hell around here....

[cytg111]

** all your pointers are pointing .. back in your face .. It's the forums way of saying, enough with flame** allready, god**it.

[skipsneeky2]

Wonder how long it will be before 6 core comes with a mainstream price in the $200-$300 range?

I suspect when the time for a mainstream affordable 6 core comes out from intel,they might release a low stockclocked model perhaps on a new stepping without ht and sell it for about $300 or so.

Sorta might happen like when the g0 stepped e6750 and q6600 came out.

[minitron]

Originally Posted by AtenRa

Intel XEON E5 takes 91.5 seconds to complete the run
FX8350 needs 116 seconds to complete the run

FX8350 needs 24,5 seconds more to complete then run

116 seconds are ~26,7% more than 91,5 seconds

Edit : That makes the FX8350 ~26,7% slower than Intel XEON E5 because it needs 24,5 seconds more to finish the run.

Intel XEON E5 needs ~21,1% less time to complete the run than FX8350

That makes Intel XEON E5 ~21,1% faster than FX8350.

More bad math.

The i7 completing the task in 79% of the time it would take the FX which is not the same as the i7 being 21% faster.

I guess I must use an example for AMD fanboy:
Assuming constant speed and no acceleration...
Car A travels 1/4 Mile in 12 seconds
Car B travels 1/4 Mile in 24 seconds

Car A is not 50% faster than Car B (12/24)
Car A is 100% faster than Car B (24/12)

Or you can just try to do the reverse equation (for your terrible math) in which case 91.5 x 1.21 =/= 116.

Edumacations, go gets thems?!?

[inf64]

You cannot get it,give up man.It's just over your head...

[Abwx]

Originally Posted by minitron

The i7 completing the task in 79% of the time it would take the FX which is not the same as the i7 being 21% faster.

So if the 8350 is at 100% and the i7 as in your exemple at 79% ,
how much less time in % does is take to do the task ????

When looking at the faster CPU we start from 116 , that is
from the slower to the faster , so 116 will be the base , hence
the i7 will be 24.5/116 x 100 = 21.1% faster.

When computing by how much the slowest is slower , we go from 91.5 to 116 , that is, from faster to slower , as such the faster, 91.5 , will be the base , this time the 8350 will be 24.5/91.5 x 100 = 26.77% slower....

Inf64 is right , you understand nothing at basic maths
since you dont even realise that the faster realize the lowest
absolute number while the slowest do the higher absolute number.

Originally Posted by inf64

You cannot get it,give up man.It's just over your head...

I sense some uncontrolled convolution in an imaginary domain...

[minitron]

The i7 completes the task in 91.5s
The FX completes the task in 116s

The i7 is not 21% faster than the FX [1 - (91.5/116)]
The i7 is 26% faster than the FX (116/91.5)

I understand there's 10 numbers and it's hard for some of you to master them all but it's something you'll probably have to learn eventually. Then again maybe not.

Take a book out of inf64's playbook: "I'm done with you."