First complete review of Haswell i7-4770K

[inf64]

If the average is 20 percent different, you cant suddenly make that 40% in one game without some other game then being equal. Otherwise the math does not add up.

That aside, who cares anyway, any of the 3 igps will still be lower in performance than a HD7750, so you are still in limbo of better than good enough for non-demanding uses, but barely adequate (at best) for gaming.

I am somewhat disappointed in the lack of cpu improvement in Haswell, but I could care less about the igp.

Believe it or not people care about iGPU performance and OEMs and intel know about that. It's all about cutting costs on the whole machine price and having iGPU means -50$ on the total price. For us maybe 50$ is not much but for a lot of buyers on a budget it might be.

 

[mikk]

@ mikk

So you showed us that 1 year old 5800K has noticeably better iGPU than yet to launch 4770K with HD4600 inside. Richland will offer +15% so yes it will walk away (as I mentioned Richland in my post ). 1.15x1.07=1.23 or 23% better performance on average for just clock bumped Trinity (which is what Richland is). There will be games where Trinity's +20% will turn into +40% on Richland. AMD is positioned nicely on desktop I have to say.

Richland will offer not more than 5-10%. GPU frequency basically the same. Even a 30% advantage would be a real shame for AMD against Intels GT2. Once again you were wrong about your 3dmark statement.

 

[Enigmoid]

@ mikk

So you showed us that 1 year old 5800K has noticeably better iGPU than yet to launch 4770K with HD4600 inside. Richland will offer +15% so yes it will walk away (as I mentioned Richland in my post ). 1.15x1.07=1.23 or 23% better performance on average for just clock bumped Trinity (which is what Richland is). There will be games where Trinity's +20% will turn into +40% on Richland. AMD is positioned nicely on desktop I have to say.

Performance increases are on mobile NOT desktop. Mobile has problems maintaining turbo because of thermals and throttles under heavy load (this can already be seen in amd's pre-release performance slides where the low voltage mobile parts got a greater % increase than the SV parts). Desktop doesn't have this problem and the a10-5800k can keep its full turbo under load.

a10-5800k -3.8 ghz base, 4.2 turbo, graphics at 800 mhz

a10-6800k-4.1 bas, 4.4 turbo, graphics at 844 mhz.

Where is 15% there?

 

[AtenRa]

@mikk

Still waiting for the link or settings for the table you posted

 

[krumme]

Richland will offer not more than 5-10%. GPU frequency basically the same. Even a 30% advantage would be a real shame for AMD against Intels GT2. Once again you were wrong about your 3dmark statement.

Unlike some of you guys, i have actually tried to play real games using hd3000 and hd4000 - and not for the purpose of benchmarking. And its bad, lacking picture quality at same setup compared to low end NV & AMD, and minimum framerates that is bad. And then for hd3000 this kind of chopping picture playing most hd mkv files.

The last 10 years i think drivers from AMD and NV have been fine. Its nothing like the experience from the 90 with crashing all over. I find the driver discussion mood for that point, as they always work in my simple cases, but Intels still quite behind here. Intel both need to have a new arch for sure because of efficiency for especially the mobile market and mm2/perf. But they really need to stop thinking only like they are producing cpu, and start allocating some serious ressources for driver development. Big time.

I take and NV or AMD card that seems to be 10% or 20% slower than any given Intel solutions, benchmarked by Anandtech for the usual bunch of games, any day of the week.

I will get this GT2, and next time someone gives the usual bs about how fantastic it is compared to any nv or amd solution, they have to be forced to play only using that gpu.

 

[galego]

Neither of those articles do justice to Amdahl's Law, it is no wonder the conclusions lend themselves to further incorrect assertions.

Amdahl's Law has never been "an argument against massively parallel processing"...that is a strawman position concocted to then dispute and refute Amdahl's Law on a falsely asserted premises.

Amdahl's Law is a limit, it tells you where you must take your code, at a minimum, if you desire to have a given speedup on your computations.

Amdahl's Law has been used in the past as "a fatal limit to the usefulness of parallelism". Some people has believed that this law introduces some fundamental limit to parallel computation. They were wrong as cited articles note.

Amdahl's Law provides a limit for code that fits in the assumptions made to derive the law. There exist code that does not follow that law.

 

[AtenRa]

OK here is the link (translated) for the review

http://translate.google.com/transla...1&u=http://www.itocp.com/htmls/33/n-4533.html

The CPUs are ES.

Just to point out that they have used the latest Intel GPU driver but only the Catalyst 13.2 Beta7 for the Trinity A10-5800K and not the latest 13.4 or 13.5Beta. Also all of the Gaming benchmarks were run at Low 720p except for DIRT Showdown that was run at Low 1080p.

Also, even they had a 2133MHz memory they run it at 1600MHz

 

[galego]

OK here is the link (translated) for the review

http://translate.google.com/translate?sl=zh-CN&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&eotf=1&u=http%3A%2F%2Fwww.itocp.com%2Fhtmls%2F33%2Fn-4533.html

The CPUs are ES.

Just to point out that they have used the latest Intel GPU driver but only the Catalyst 13.2 Beta7 for the Trinity A10-5800K and not the latest 13.4 or 13.5Beta. Also all of the Gaming benchmarks were run at Low 720p except for DIRT Showdown that was run at Low 1080p.

Also, even they had a 2133MHz memory they run it at 1600MHz

Yes. Subtract about a 5-10% performance from running the AMD APU with memory below its stock speed.

 

[Charles Kozierok]

Amdahl's Law has been used in the past as "a fatal limit to the usefulness of parallelism". Some people has believed that this law introduces some fundamental limit to parallel computation. They were wrong as cited articles note.

Amdahl's Law provides a limit for code that fits in the assumptions made to derive the law. There exist code that does not follow that law.

Sorry, but this doesn't make any sense. It's like saying that Newton's law of universal gravitation is "wrong" because we can shoot rockets into space.

Once again, Amdahl's Law is not about parallelism, regardless of what "some people" may or may not believe. It doesn't introduce any "fundamental limit" to parallel computation. It doesn't introduce any limits at all -- it is a simple mathematical expression of reality under specific conditions, and by definition only applies to those conditions, but they describe the vast majority of computing problems. The only place that Amdahl's Law "breaks" is the special case where the workload can expand indefinitely.

 

[SiliconWars]

Also, even they had a 2133MHz memory they run it at 1600MHz

That's just sad.

 

[mikk]

Trinity officially support DDR3-1866 not DDR3-2133. The gap narrows down, bad news for AMD notebooks. Haswell GT2 should be enough and GT3 just walk away easily.

 

[inf64]

GT3 just walk away easily.

Yep, price wise it will walk away most definitely .

 

[frozentundra123456]

Does anybody buy a top end quad core cpu to run it on the IGP? Really?????

 

[galego]

Sorry, but this doesn't make any sense. It's like saying that Newton's law of universal gravitation is "wrong" because we can shoot rockets into space.

It is more like saying that Newton's law of universal gravitation is "wrong" because we have GPS.

Once again, Amdahl's Law is not about parallelism, regardless of what "some people" may or may not believe. It doesn't introduce any "fundamental limit" to parallel computation. It doesn't introduce any limits at all -- it is a simple mathematical expression of reality under specific conditions, and by definition only applies to those conditions, but they describe the vast majority of computing problems. The only place that Amdahl's Law "breaks" is the special case where the workload can expand indefinitely.

Here you essentially repeat what some of us have said and what is said in the two links that I gave before.

Trinity officially support DDR3-1866 not DDR3-2133.

And as noted before, the review did run memory @ 1600 MHz instead of @ the proper stock speed.

 

[Charles Kozierok]

OK here is the link (translated) for the review

Once again.. who are these people, and why should I care about what they have to say on this or any other subject?

 

[Blandge]

Unlike some of you guys, i have actually tried to play real games using hd3000 and hd4000 - and not for the purpose of benchmarking. And its bad, lacking picture quality at same setup compared to low end NV & AMD, and minimum framerates that is bad. And then for hd3000 this kind of chopping picture playing most hd mkv files.

The last 10 years i think drivers from AMD and NV have been fine. Its nothing like the experience from the 90 with crashing all over. I find the driver discussion mood for that point, as they always work in my simple cases, but Intels still quite behind here. Intel both need to have a new arch for sure because of efficiency for especially the mobile market and mm2/perf. But they really need to stop thinking only like they are producing cpu, and start allocating some serious ressources for driver development. Big time.

I take and NV or AMD card that seems to be 10% or 20% slower than any given Intel solutions, benchmarked by Anandtech for the usual bunch of games, any day of the week.

I will get this GT2, and next time someone gives the usual bs about how fantastic it is compared to any nv or amd solution, they have to be forced to play only using that gpu.

A friend of mine played all the way through Skyrim with HD4000 (3770K) and said he had no problems at all.

 

[Idontcare]

Amdahl's Law has been used in the past as "a fatal limit to the usefulness of parallelism". Some people has believed that this law introduces some fundamental limit to parallel computation. They were wrong as cited articles note.

Amdahl's Law provides a limit for code that fits in the assumptions made to derive the law. There exist code that does not follow that law.

This is just more strawman argumentation.

There is no such thing as making code that doesn't follow the law, all you've done is make code that has different parameters that define a different limiting value.

I find this whole vector of "must debunk Amdahl" laughable, as if there is anything to debunk and as if debunking Amdahl actually accomplishes anything. Reminds me so much of the so-called "1GHz barrier". Bunch of nonsense.

Amdahl's law will only fail to make any sense if the individual attempting to apply Amdahl's law has failed themselves to properly characterize the workload and application in question.

Do your part of the job correctly and the math naturally falls into place, don't do it correctly and suddenly you find yourself grasping for straws and magic to explain your seemingly "Amdahl Law busting!" data set.

Does anybody buy a top end quad core cpu to run it on the IGP? Really?????

:whiste: (looks around) I know I am a corner-case example, but one of my 3770k's is doing exactly that. Mind you I don't game on it, but I do use the IGP to power the desktop and 2D apps for a compute box that I remote-desktop into.

 

[SiliconWars]

Trinity officially support DDR3-1866 not DDR3-2133. The gap narrows down, bad news for AMD notebooks. Haswell GT2 should be enough and GT3 just walk away easily.

Richland is 2133 MHz capable and will gain 10%+ in almost every game on the increased bandwidth alone. With higher cpu clocks (cpu performance is also a limiter at lower resolutions) and 5% gpu clocks, 20% higher fps should be expected on most titles.

We don't know if GT3 will run into heavy bandwidth limitations as well but I would assume so. Then you have to factor the dual core with HT loses a lot on true quad core performance in many games.

If you mean the i5's with GT3 then it doesn't matter as they'll cost at least double.

 

[AtenRa]

:whiste: (looks around) I know I am a corner-case example, but one of my 3770k's is doing exactly that. Mind you I don't game on it, but I do use the IGP to power the desktop and 2D apps for a compute box that I remote-desktop into.

You are not the only one, a lot of people need/want the CPU performance of the 3770/K without needing a dGPU.

 

[frozentundra123456]

Richland is 2133 MHz capable and will gain 10%+ in almost every game on the increased bandwidth alone. With higher cpu clocks (cpu performance is also a limiter at lower resolutions) and 5% gpu clocks, 20% higher fps should be expected on most titles.

We don't know if GT3 will run into heavy bandwidth limitations as well but I would assume so. Then you have to factor the dual core with HT loses a lot on true quad core performance in many games.

If you mean the i5's with GT3 then it doesn't matter as they'll cost at least double.

Maybe for someone that builds their own system. But do you think OEMs are going to put expensive ram in budget apu box? Heck, sometimes they still dont even put in dual channel.

 

[NTMBK]

Does anybody buy a top end quad core cpu to run it on the IGP? Really?????

If the IGP is powerful enough, sure. Right now that category only encompasses power constrained environments like HTPCs and laptops, of course.

 

[Idontcare]

You are not the only one, a lot of people need/want the CPU performance of the 3770/K without needing a dGPU.

Possible thread derail, and I apologize, but I wish my FX8350 was an APU for the same reason. Love the CPU horsepower, but the fact that I have to throw in a discrete GPU card that is going to suck down another 30W idle for 2D desktop stuff just bums me out.

Hopefully steamroller powered APUs will have FX SKUs so people don't have to step back from so much CPU power just to get an IGP with their AMD rigs as they do now with the Trinity vs Piledriver FX SKUs.

 

[AtenRa]

Trinity officially support DDR3-1866 not DDR3-2133. The gap narrows down, bad news for AMD notebooks. Haswell GT2 should be enough and GT3 just walk away easily.

I can run the memory at 2133MHz on my A10-5800K without OC on the Asrock FM2A85X-iTX.

Just for reference, A10-5800K Default with 2133MHz ram, catalyst 13.4 on Win 8 Pro 64bit.

 

[Enigmoid]

Does anybody buy a top end quad core cpu to run it on the IGP? Really?????

Many (probably most) OEM systems ship a 3770/3770k and run the igp.

Anyone who needs brute force cpu grunt finds its useful also.

Trinity officially support DDR3-1866 not DDR3-2133. The gap narrows down, bad news for AMD notebooks. Haswell GT2 should be enough and GT3 just walk away easily.

Plus the fact that a $500 trinity notebook runs either 1333 mhz or 1600 mhz ram (high latency) sometimes in single channel. Even high end i7 laptops generally don't have better than 1600 mhz cl 11 ram.

Considering how fast gt2 on the desktop is and how it compares to the a10-5800k on the mobile side it should be very close to mobile trinity/richland (amd reduces the mobile igp clocks more than intel).

 

[mikk]

And as noted before, the review did run memory @ 1600 MHz instead of @ the proper stock speed.

That doesn't make a big difference. Not more than 5-10%. inf64 underestimated Haswell GT2 or overestimated AMD based on wrong 3dmark assumptions. Who would have thought that there is such a small gap between Haswell GT2 and A10-5800k. I think their new driver plays a big role. 18% from the driver and 25-30% from Haswell GT2 and we get a 45-50% jump from last year.