Intel Core i7 4770K "Haswell" Benchmarks On Ubuntu Linux

[YAFU]

Why is the FX8350 mysteriously missing from some of the benchmarks?
nevermind, I'm retarded.

Because PhoronixTestSuite can only measure up to two-digit numbers in temperature and power consumption... Or maybe the FX-8350 just has exploded (just kidding AMD guys).

Some tests are about the iGPU, and the FX-8350 does not have integrated GPU. In the other I have no idea why.

[Kivada]

Yes, AMD has come back considerably lately (which I love to see). But keep in mind that the FX lines doesnt have integrated graphics, so you need to buy some graphic card to have something comparable. Plus, the Intel chip has a lot more OC'ing potential because it has a much saner TDP. In terms of graphics, you need to compare with an amd APU, and those are a lot slower for CPU usage (even the i5's beat them)

Cheers!

Ok, then let's do a price comparison here, what the best bang for your computing buck? I'm using the same very popular shop for all parts.

$179.98 basic A10-5800K system

$264.98 basic FX-8350 system

$414.98 basic i7-4770K system

I'm assuming that ram, psu, drives and case are identical across all systems and thus left out of the comparison.

[smitty3268]

Wierd. I do scientific heavy threaded stuff and my testing showed 20% performance improvement on core i7 with HT enabled over disabled. And yes it is very cpu heavy and I burn up core i7 laptops (the RAM isn't throttled and causes the hard crash).

I am curious how amd's new architecture might handle things but since we're so heavy on double precision I'm not confident we would get good performance. Our software running on core i7 runs circles cpu wise around the pre bulldozer opterons (dual socket sandy-e easily outruns quad socket opteron, about 2x faster on cpu heavy loads).

I noticed on other reviews how the 3930k still outruns this new haswell on heavily threaded loads.

Early hyper-threading hardware (like P4) was famous for sometimes slowing down applications rather than speeding them up, but that hasn't been an issue for a while.

These days, the bad apps mostly just don't see any speed increase from HT, but won't actually be any worse. There's always one or 2 exceptions, of course, but for the most part you can consider it a solved problem. A lot of that is probably due to OS patches, too, to inform the scheduler to use the primary cores and not try to stick too much work on the HT virtual cores.

AMD's architecture is somewhat similar in that they also share hardware resources between multiple threads on a core, but with AMD much less is shared and blocking the other thread, which means it should scale better and provide better performance when going over 4 threads than an Intel 4 core with HT processor. That said, it's primarily for integer code, so i'm not sure how much of a gain you'd get from your double precision fp heavy codebase.

[Kivada]

Yes heat is going up like a rocket! /sarcasm

the a10-5800k is 3.8 ghz base with a TDP of 100w the a10-6800k is going to be 4.1 ghz base with a TDP of 100W

So how is heat going up from boosting clocks?

Seriously the 8350 is only 125w the 4770k is 84 if you can make a GOOD argument about why 41 watts is a HUGE difference on a high end desktop please let me know. (And please don't say it'll save money)

With leaving a PC on 24/7 the cost difference due to the CPU power draw is 31.25 a year, let's see 150 / 31.25 = 4.8. Congratulations your over priced intel crap will pay for itself in 4.8 YEARS!

AMD and Intel state their TDP differently, Intel goes with the average while AMD goes with maximum. Thus AMD CPUs have always had higher TDP on paper but in practice Intel systems usually ran considerably hotter. It doesn't help Intel at all that they go very cheap on the stock cooling. I've seen many an Intel box throttled to slow as molasses speeds due to dust buildup on their dinky blocks with the worst HSF retention method ever conceived. A both through kit is pretty much required as well since they glue on the IHS to the CPU the temps are very high, requiring you to remove it if you want to have a decent stable overclock, I've seen this done with razor blades and hammers...

[YAFU]

I'm assuming that ram, psu, drives and case are identical across all systems and thus left out of the comparison.

If we assume that the AMD has higher power consumption, I think in some cases it may need a higher PSU power.
At least in Asus you have very cheap motherboards with PCI-e 3.0 support for intel (assuming you want to use a discrete card). The cheap Asus motherboards for AMD FX I've seen do not support PCI-e 3.0.

I think each has its advantages and disadvantages. I hope that there is always competition between AMD and Intel. It would not be good for consumers that there is much difference between the two, and that only one dominates the market.

[Kivada]

If we assume that the AMD has higher power consumption, I think in some cases it may need a higher PSU power.
At least in Asus you have very cheap motherboards with PCI-e 3.0 support for intel (assuming you want to use a discrete card). The cheap Asus motherboards for AMD FX I've seen do not support PCI-e 3.0.

I think each has its advantages and disadvantages. I hope that there is always competition between AMD and Intel. It would not be good for consumers that there is much difference between the two, and that only one dominates the market.

I don't know about you, but I never cheap out on the PSU, it's the last line of defense your system has in case of a power surge. I'd rather replace the $100-130 PSU high efficiency, low ripple, very good standby efficiency and with capacity to spare to compensate for the aging of the capacitors because it was properly designed and took the hit to save the box instead of some el cheapo PSU that takes out half of your hardware.

PCIe2.0 isn't even being fully saturated. PCIe3.0 is a solution looking for a problem, it may only make minuscule difference in top end GPUs 2 generations from now. Test it yourself, get some tape out and tape over the lanes and turn your 16x card into an 8x, 4x and 1x card and run the same GPU benchmarks on it every time.

[duby229]

because in the last bench the 8350 would have won, so naturally it was excluded.

[duby229]

AMD and Intel state their TDP differently, Intel goes with the average while AMD goes with maximum. Thus AMD CPUs have always had higher TDP on paper but in practice Intel systems usually ran considerably hotter. It doesn't help Intel at all that they go very cheap on the stock cooling. I've seen many an Intel box throttled to slow as molasses speeds due to dust buildup on their dinky blocks with the worst HSF retention method ever conceived. A both through kit is pretty much required as well since they glue on the IHS to the CPU the temps are very high, requiring you to remove it if you want to have a decent stable overclock, I've seen this done with razor blades and hammers...

Also Intel doesnt (or didnt) count the integrated northbridge in the TDP figures either, so memory controller and the likes don't get factored into TDP either.

[YAFU]

I don't know about you, but I never cheap out on the PSU

In my previous computers I've spent lots of money on PSU, and two times the PSU are broken!
So in my last computer I said: What the hell! and I bought a cheap "Sentey" case with PSU included! LoL
Anyway I have a one year warranty on all computer....

[smitty3268]

I don't know about you, but I never cheap out on the PSU, it's the last line of defense your system has in case of a power surge. I'd rather replace the $100-130 PSU high efficiency, low ripple, very good standby efficiency and with capacity to spare to compensate for the aging of the capacitors because it was properly designed and took the hit to save the box instead of some el cheapo PSU that takes out half of your hardware.

PCIe2.0 isn't even being fully saturated. PCIe3.0 is a solution looking for a problem, it may only make minuscule difference in top end GPUs 2 generations from now. Test it yourself, get some tape out and tape over the lanes and turn your 16x card into an 8x, 4x and 1x card and run the same GPU benchmarks on it every time.

PCIe3.0 does improve some compute benchmarks, just not graphics.