FX 8370 Review

[AtenRa]

hmmm am i missing something here?

it seems that amd cards are doing better here. I dont get your bolded claim.

He means for the CPU, not the GPU. I would wait for a new more optimized Dead Rising 3 driver or patch.

 

[Fox5]

He means for the CPU, not the GPU. I would wait for a new more optimized Dead Rising 3 driver or patch.

AMD vs...

Nvidia

As you scale the graphical settings, the game starts to become GPU limited, so the high end Intel cpus will perform better with the faster Radeon cards. At lower settings, Nvidia has far less of a cpu bottleneck.
This is also true in Battlefield 4 for DX11, but that game has Mantle for AMD users.

I think the same thing also held true in Civilization V, due to its support for DX11 multithreading. I think AMD implemented a one game solution where they will do DX11 multithreading, but just in Civ V. Since there's only about 3 games that even support DX11 multithreading, I suppose they could do that again.

 

[frozentundra123456]

Are those resolutions correct? I thought they tested everything at 1080p. And the CPU core loading data shows a resolution of 1920 x 1080.

 

[Fox5]

Are those resolutions correct? I thought they tested everything at 1080p. And the CPU core loading data shows a resolution of 1920 x 1080.

I'd guess they lowered the resolution to show a cpu bottleneck.

 

[bunnyfubbles]

AMD vs...

Nvidia

As you scale the graphical settings, the game starts to become GPU limited, so the high end Intel cpus will perform better with the faster Radeon cards. At lower settings, Nvidia has far less of a cpu bottleneck.
This is also true in Battlefield 4 for DX11, but that game has Mantle for AMD users.

I think the same thing also held true in Civilization V, due to its support for DX11 multithreading. I think AMD implemented a one game solution where they will do DX11 multithreading, but just in Civ V. Since there's only about 3 games that even support DX11 multithreading, I suppose they could do that again.

a bit off tangent, but I still don't know how you could come to the conclusion that the game is "well threaded" when the 6 core i7 comes in second place to 4 core chips... or how the i3 dual core Sandy chip is faster than all the AMD CPUs except for the two heavily clocked Piledriver chips (but only in the nVidia test is the i3 2100 slower than the 9000s)... or how the "8 core" 8150 and 8350 have almost no advantage over their "4 core" 4100 and 4300 counterparts... or how performance is scaling almost perfectly with clock rate...

 

[Ramses]

Is anyone really surprised by any of this? This all looks like pretty much any other game benchmark I've seen for the last three or four years at least. When I see Intel CPU's being used for benchmarks(and they always are), I just subtract 15% or so for what my system would do with the same GPU. I can remember last time I was into games the 9800pro was a hot card, and nvidia drivers were better back then and much discussion was made about what games favored what drives/cards. It was all pretty much the same bit.
And it still looks like my 8350 and 280x crossfire junk will play dead rising ok.
Still no regrets.. https://www.youtube.com/watch?v=sGTXqotOUqA

 

[escrow4]

Is anyone really surprised by any of this? This all looks like pretty much any other game benchmark I've seen for the last three or four years at least. When I see Intel CPU's being used for benchmarks(and they always are), I just subtract 15% or so for what my system would do with the same GPU. I can remember last time I was into games the 9800pro was a hot card, and nvidia drivers were better back then and much discussion was made about what games favored what drives/cards. It was all pretty much the same bit.
And it still looks like my 8350 and 280x crossfire junk will play dead rising ok.
Still no regrets.. https://www.youtube.com/watch?v=sGTXqotOUqA

Because your FX will dip heavily and frequently, that hitching is a nightmare. Intel is butter smooth. I don't game on a PC for "OK". I game to push it the limit settings wise and 50-60FPS solid is mandatory. Otherwise I may as well buy a console. The 4770 non K I have in my gaming box loads up to 3.7GHz or so turbo in game and its smooth. Period.

 

[Fox5]

a bit off tangent, but I still don't know how you could come to the conclusion that the game is "well threaded" when the 6 core i7 comes in second place to 4 core chips... or how the i3 dual core Sandy chip is faster than all the AMD CPUs except for the two heavily clocked Piledriver chips (but only in the nVidia test is the i3 2100 slower than the 9000s)... or how the "8 core" 8150 and 8350 have almost no advantage over their "4 core" 4100 and 4300 counterparts... or how performance is scaling almost perfectly with clock rate...

The game is well threaded because I can see is pegging all 6 cores at times, and consistently pegging 3-4 of them.
However, a game can be well threaded, and still have all its threads waiting on a single thread to complete. The implication I get from those charts is that everything is waiting on the render thread, which is why nvidia's better threaded drivers double performance on AMD, but only give a 50% increase on Intel as the bottleneck moves elsewhere.

 

[frozentundra123456]

I pretty much see it that way as well. That is why I still don't buy the argument that FX is going to catch up or surpass Intel quads overall as games become more well threaded. In some games, they are competitive, but in other games that also use a lot of threads it seems like fast performance per core is also important.

 

[AtenRa]

Now have a look how a well threaded game utilizes the CPU cores.

 

[Abwx]

Floating performances with PoVray tracer, FMA looks to be enabled for the CPUs that support it according to the soft sites, this is likely the case given the ST scores.

http://hothardware.com/Reviews/AMD-FX8370-and-FX8370E-8Core-CPU-Reviews/?page=4

 

[Phynaz]

Wonder why a 4770k or 4790k isn't in that chart...

 

[Ramses]

Because your FX will dip heavily and frequently, that hitching is a nightmare. Intel is butter smooth. I don't game on a PC for "OK". I game to push it the limit settings wise and 50-60FPS solid is mandatory. Otherwise I may as well buy a console. The 4770 non K I have in my gaming box loads up to 3.7GHz or so turbo in game and its smooth. Period.

I'm sure it does but I don't notice. I'm glad we have the option of buying chips that don't dip and such, but I spend maybe 3% of my screen time gaming. Still no regrets..
In fact I think AMD is going to get another $250 of my money for the 9590 this week, I just noticed 8350's are pulling $140 or so on ebay so the outlay is pretty minimal. Gotta get my temps up, winter's coming..


If anyone wants to get rich btw, find a way to plug a trackball and keyboard and remap keys on a console. I'd be thrilled to not have to jerk around with buggy drivers and patches and this and that gaming on a PC, but my first and only console was an Atari, and I can't stand those little controllers. (and I'd be even MORE content with AMD CPU's )

 

[Enigmoid]

Why does the i5-4670k have such poor scaling?

 

[Abwx]

Why does the i5-4670k have such poor scaling?

Because of the turbo on ST..?.

Edit : Surely a bios issue, a 4570 scores 1271, a 3770k does 1357 and surprisingly the FX8150 of old scores 1447 wich put it at a better IPC than the 3770k, there s about no doubt that FMA is used given all thoses scores, particularly the low or any advantage of the IB i7 compared to HW i5.

http://www.computerbase.de/2014-09/amd-fx-8370e-im-test/2/#diagramm-pov-ray

There s povray scores on the meddle charts..

 

[jhu]

Why does the i5-4670k have such poor scaling?

Higher turbo clocks for single thread vs. full threads. You'll find IPC/core is still the same.

 

[jhu]

Because of the turbo on ST..?.

Edit : Surely a bios issue, a 4570 scores 1271, a 3770k does 1357 and surprisingly the FX8150 of old scores 1447 wich put it at a better IPC than the 3770k, there s about no doubt that FMA is used given all thoses scores, particularly the low or any advantage of the IB i7 compared to HW i5.

http://www.computerbase.de/2014-09/amd-fx-8370e-im-test/2/#diagramm-pov-ray

There s povray scores on the meddle charts..

Povray does not use FMA. The issue, I suspect, is that Povray for Windows is compiled with MSVC. Also, it's running on Windows. Things look a bit different with gcc/icc on Linux

 

[Fox5]

Now have a look how a well threaded game utilizes the CPU cores.

How does a well threaded game use its cores?

This one looks like 6 threads at 100% for intense parts, and 6 threads at 50% for less intense parts.

 

[davie jambo]

Because your FX will dip heavily and frequently, that hitching is a nightmare. Intel is butter smooth. I don't game on a PC for "OK". I game to push it the limit settings wise and 50-60FPS solid is mandatory. Otherwise I may as well buy a console. The 4770 non K I have in my gaming box loads up to 3.7GHz or so turbo in game and its smooth. Period.

I assure you DR3 on intel is not "butter smooth" , my i5 there are frequent drops to the mid 30s

His FX will be fine just like they are fine for every single PC game

 

[Ramses]

Glad I didn't buy the I5 I looked at for about ten seconds this afternoon..

Course I didn't buy the 9590 yet either since nobody but provantage is showing new pricing yet.

 

[davie jambo]

Glad I didn't buy the I5 I looked at for about ten seconds this afternoon..

Course I didn't buy the 9590 yet either since nobody but provantage is showing new pricing yet.

If your gaming you'd be as well as buying an i5 tbf

I like AMD but you'd be as well as buying haswell i5

 

[SlowSpyder]

Hey, look. Another game that my CPU has more than enough horsepower to play well (as long as there is enough GPU) while people on a forum discuss how poorly the FX performs. Different day, same thing as always.

I'd like to see the E CPU's benched in this game with their lower base clock.

 

[SPBHM]

Hey, look. Another game that my CPU has more than enough horsepower to play well (as long as there is enough GPU) while people on a forum discuss how poorly the FX performs. Different day, same thing as always.

I'd like to see the E CPU's benched in this game with their lower base clock.

if your target is 60FPS and you have a Radeon, apparently not

while a much older and cheaper 2500K OC apparently would do quite well in this situation!?

there is a 3970X in there, I don't think the 5960X would be any slower.

 

[davie jambo]

if your target is 60FPS and you have a Radeon, apparently not

while a much older and cheaper 2500K OC apparently would do quite well in this situation!?

there is a 3970X in there, I don't think the 5960X would be any slower.

i5 2500k & R9 280x here

Frames go from 30-90 , probably about the same as the FX

 

[SlowSpyder]

if your target is 60FPS and you have a Radeon, apparently not

while a much older and cheaper 2500K OC apparently would do quite well in this situation!?

there is a 3970X in there, I don't think the 5960X would be any slower.

Seems to be a driver issue with the Radeon. Both the 2500k and FX CPU's see a lot less performance on the R9 290X compared to the 780TI. To be fair, if your target is no less than 60FPS, it looks like with current drivers you won't get it with even a 4770k either, if you have a Radeon.