AMD will launch AM4 platform in March 2016 says industry source

[ShintaiDK]

Moving the goalposts? I know that global warming prevented building a snowman instead. Before jumping to conclusions, it would help to check, how many cores the game utilizes and what after gfx the next heavy thread is doing regarding CPU utilization.

Lucky they made CPU scaling tests too. "Moar cores" isn't going to benefit either so far.

So current DX12 experience so far tells us:
ST performance rules.
4 threads is still more than plenty.

With faster CPUs you can clearly see ~50% and ~80% improvements depending on detail settings. As long as the already better multithreaded Nvidia drivers don't work on AMD cards, it would be stupid to say, that DX12 won't help them based on an arbitrarily chosen CPU bottleneck to hide the effect.

Its not something AMD couldn't have fixed in their DX11 drivers.

 

[VR Enthusiast]

2 demos. And then devs will use the CPU power for something else, leaving AMD in the hole again.

AMD CPUs can easily match Intel when all CPU resources are used.

Lucky they made CPU scaling tests too. "Moar cores" isn't going to benefit either so far.

So current DX12 experience so far tells us:
ST performance rules.
4 threads is still more than plenty.

The only thing we can tell from Ashes is that nobody seems to know what is going on there.

We know from many other benchmarks that when the FX cores are being used then it is much faster than the i3 and i5, and often equals the quad i7s.

With properly-done DX12, a couple of node shrinks and 40% IPC uplift there is no logic behind assuming that Octo-core Zen (maybe even 6 cores with SMT) won't be at least as good as a quad i7. Even now we see well-threaded DX11 games showing gains for 12 threads vs 8 threads vs 4.

 

[dark zero]

Lucky they made CPU scaling tests too. "Moar cores" isn't going to benefit either so far.

So current DX12 experience so far tells us:
ST performance rules.
4 threads is still more than plenty.



Its not something AMD couldn't have fixed in their DX11 drivers.

So DX 12 is trash then since they were focused on MT and at 6 cores.
Seems that developers are even lazzier than the ones from the crash of 1983.... hope that another crash happen if that so in order to purge the gaming industry.

 

[tential]

So DX 12 is trash then since they were focused on MT and at 6 cores.
Seems that developers are even lazzier than the ones from the crash of 1983.... hope that another crash happen if that so in order to purge the gaming industry.

Please quote where DX12 is focused at 6 cores (A SMALL segment of the market).

That would be an insanely poor move....

Seems like you just don't know the focus of DX12....

 

[tential]

Lucky they made CPU scaling tests too. "Moar cores" isn't going to benefit either so far.

So current DX12 experience so far tells us:
ST performance rules.
4 threads is still more than plenty.



Its not something AMD couldn't have fixed in their DX11 drivers.

No matter what, there will always be an external reason why AMD CPUs/GPUs have issues that people blame on other companies. It's NEVER AMD's fault...

 

[Dresdenboy]

Lucky they made CPU scaling tests too. "Moar cores" isn't going to benefit either so far.

So current DX12 experience so far tells us:
ST performance rules.
4 threads is still more than plenty.



Its not something AMD couldn't have fixed in their DX11 drivers.

As modern multithread approaches use flexible task handling across cores, fewer but much stronger cores work at least as good. But one heavy task might be enough to give the stronger cores an advantage. PD surely isn't one of them. Agreed so far.

Regarding DX11: This is an economical question. It's a MS API, likely causing a couple of man years to get fixed this way and delaying other work, while the majority of AMD cards runs on Intel systems where not as much could be gained. Without checking I assume that APU iGPUs are also slow enough to hide the DX11 bottleneck somewhat. So spend M$ and sparse human ressources? I wouldn't.

 

[Dresdenboy]

6 core Zen might be unlikely, as this spot could be populated by 8C w/o SMT (or 4C+SMT instead of 6C w/o SMT). Except defects manage to hit the small unrepairable cores often enough.

 

[dark zero]

Please quote where DX12 is focused at 6 cores (A SMALL segment of the market).

That would be an insanely poor move....

Seems like you just don't know the focus of DX12....

We saw the targets of DX12, if this is more of the same, we get nothing good. And that is definately a bad move from MS.

And also, if the market don't want to move, you have to force them.
Look to Apple.

 

[tential]

We saw the targets of DX12, if this is more of the same, we get nothing good. And that is definately a bad move from MS.

And also, if the market don't want to move, you have to force them.
Look to Apple.

So DX 12 is trash then since they were focused on MT and at 6 cores.
Seems that developers are even lazzier than the ones from the crash of 1983.... hope that another crash happen if that so in order to purge the gaming industry.

Please quote where DX12 is focused at 6 cores (A SMALL segment of the market).

That would be an insanely poor move....

Seems like you just don't know the focus of DX12....

Again, where do you think DX12 is meant to focus on 6 cores....

Like I said, do you have ANY source for your claim that DX 12 is focused at 6 cores or that is the goal of DX12...

The goal of DX12 is to have a lower level API... Unless I'm mistaken....

 

[MiddleOfTheRoad]

Please quote where DX12 is focused at 6 cores (A SMALL segment of the market).

That would be an insanely poor move....

Seems like you just don't know the focus of DX12....

Well, there are a ton of chips running six threads, perhaps that is what he meant. It would make sense to dial software in for i7 for the enthusiast market.

 

[tential]

Well, there are a ton of chips running six threads, perhaps that is what he meant. It would make sense to dial software in for i7 for the enthusiast market.

Your comment makes just as little sense. Youre saying direct x 12 is targeted at 6 threads?

Did you just say you want to dial dx12 to the enthusiast market? Why is God's name would we create an api targeted at the enthusiast crowd?

Are we just making up the first story that remotely fits the scenario were trying to create to show the fx line to benefit from dx12 greatly?

Or just look at the actual facts.

Nothing can save the fx line for gaming..... Sure you can make it better but it isn't beating out current i3s....

Saying that dx12 should benefit amd and make the current fx lines better than an i3 will not make it so.....

Saying "wait for proper dx 12 titles"
Lol, that's a classic amd fan move as always thouhh for amd fans. Wait for dx12 real titles they'll show benefits in what amd is focusing on. With amd it's always waiting.

 

[MiddleOfTheRoad]

Your comment makes just as little sense. Youre saying direct x 12 is targeted at 6 threads?

Did you just say you want to dial dx12 to the enthusiast market? Why is God's name would we create an api targeted at the enthusiast crowd?

Or just look at the actual facts.

Nothing can save the fx line for gaming..... Sure you can make it better but it isn't beating out current i3s....

Saying that dx12 should benefit amd and make the current fx lines better than an i3 will not make it so.....

Saying "wait for proper dx 12 titles"
Lol, that's a classic amd fan move as always thouhh for amd fans. Wait for dx12 real titles they'll show benefits in what amd is focusing on. With amd it's always waiting.

So much FUD, I don't know where to begin.

A substantial number of mainstream Intel CPU's have 8 threads, with 4 physical cores. From broadwell in tablets and 2 in ones, Core M in laptops to Haswell i7's in desktops. So yeah, dialing software in to efficiently use at least 6 threads makes a ton of sense.

AMD chips already perform well for their respective age. The biggest problem for AMD is that they are selling 2012 era processors as new for their top line product... And Intel has produced 3 generations since the Vishera debuted.

Why make an API targeted at the enthusiast crowd? Because it it weren't for PC gaming -- the desktop PC would already be dead.

Also hate to break it to you -- but Skylake dual cores still stutter (frame time variance) more than octocore FX chips in heavy multithreaded games. You can continue to masturbate to the max frames per second for the i3 if you'd like -- and ignore the frame time variance at your own peril.
Personally, I'd rather have the smoother gameplay.

So for me, Dual Core < FX Octocore < i7 Haswell / Skylake.

70% of the CPU's I run are now Intel, but this AMD misinformation is just downright dumb.

 

[DrMrLordX]

The G3258's stuttering is well-documented. Do you have any videos of the i3-6100 + dGPU stuttering in games?

 

[MiddleOfTheRoad]

The G3258's stuttering is well-documented. Do you have any videos of the i3-6100 + dGPU stuttering in games?

Well, I haven't used an i3 6100 personally so far. I've only played with the 6700K. As for the dual core reference, I was referring to a best friend who just built a G4400 system -- and the stuttering was pretty ugly. GTA 5 is really bad. Stuttering didn't seem quite as bad as it does on my G3258, but it was definitely noticeable to me (not sure if average user would notice it as much as I did -- since I have been mostly playing games on an overclocked 4790K for past 6 months).

 

[DrMrLordX]

I'm not sure that an i3 would have the same problem. But if you have access to a 6700k and are able to disable two cores in the UEFI, maybe you could (sort of) find out the answer.

 

[AtenRa]

The G3258's stuttering is well-documented. Do you have any videos of the i3-6100 + dGPU stuttering in games?

I dont have a video for the Core i3 6300 but those two will do,

Win 10 64bit
Cat 15.11Beta

Ryse - 1080p High settings

BF4 MP 64player map - 1080p Custom 120fps Cap and Mantle

 

[DrMrLordX]

Huh. Wonder if I would notice that in a vid though . . . still interesting.

Also, a 6300 losing to an 8150 in average fps? lulz. I'd like to see a 7870k in that comparison. Or a Bristol Ridge when it comes out.

 

[AtenRa]

Huh. Wonder if I would notice that in a vid though . . . still interesting.

Also, a 6300 losing to an 8150 in average fps? lulz. I'd like to see a 7870k in that comparison. Or a Bristol Ridge when it comes out.

I dont know if you can notice that in a youtube video but i could definitely feel the difference while playing.

 

[cbn]

The G3258's stuttering is well-documented. Do you have any videos of the i3-6100 + dGPU stuttering in games?

How is the G3258's stuttering well documented? Under what conditions?

In my experience the size of the GPU matters, apparently this due to the importance of CPU to GPU balance:

http://forums.anandtech.com/showpost.php?p=37882771&postcount=15

So, if I were to switch to a worse-performing GPU, there would also be less stress on my CPU, which would alleviate spikes/stutters but give worse performance?

http://www.eurogamer.net/articles/digitalfoundry-2014-pentium-g3258-review

In the highlighted video, you can see that running the game at the high preset (that's one 'notch' down from the ultra-equivalent, very high) in combination with a GTX 760 results in a night-and-day performance differential between the i7 4790K and the Pentium. The additional fidelity in the game simulation, coupled with the immense increase in GPU set-up costs, sees the Anniversary Edition Pentium struggle horrendously to keep pace. What we're seeing here is a classic case of a lack of hardware balance: the G3258 simply can't feed the GTX 760 quickly enough to sustain a consistent frame-rate.

Now, compare and contrast with the secondary analysis, where we drop the GPU down to a far more modest GTX 750 Ti, and lower the overall quality preset to the medium level. In this case, for the most part it is the graphics card that is the bottleneck, and the overall performance level lowers the i7 advantage significantly.

Also, I liked Flapdrol's post here:

http://forums.anandtech.com/showpost.php?p=37502538&postcount=22

Pentium has run every game I tried very well.

If you're not gpu limited you have to sometimes use an fps cap though. Otherwise the high priority game threads hog both cores until the background/low priority threads can't be delayed anymore, and you get noticable stutters, or even have the game stall for a second.

Anyway, I have a gtx 670 so I don't often have to do that.

In a nutshell, get a smaller card and lower detail settings in order to smooth out gameplay.

 

[cbn]

I dont have a video for the Core i3 6300 but those two will do,

Win 10 64bit
Cat 15.11Beta

Ryse - 1080p High settings

BF4 MP 64player map - 1080p Custom 120fps Cap and Mantle

Glad to see the Bulldozer octocore do so well in the BF4 64 player.

If using a "relatively large" dGPU with low level API I will have to keep that in mind.

 

[cbn]

Well, I haven't used an i3 6100 personally so far. I've only played with the 6700K. As for the dual core reference, I was referring to a best friend who just built a G4400 system -- and the stuttering was pretty ugly. GTA 5 is really bad. Stuttering didn't seem quite as bad as it does on my G3258, but it was definitely noticeable to me (not sure if average user would notice it as much as I did -- since I have been mostly playing games on an overclocked 4790K for past 6 months).

Here is a video of a G1820 dual core and R7 240 running GTA V smoothly:

http://forums.anandtech.com/showpost.php?p=37513254&postcount=44

 

[AtenRa]

Here is a video of a G1820 dual core and R7 240 running GTA V smoothly:

http://forums.anandtech.com/showpost.php?p=37513254&postcount=44

I dont believe you can notice high frame times on youtube videos. As you can see from my slides above, even if you have high frame rate(fps) you may still have high frame times that will lead to stuttering.

 

[cbn]

I dont believe you can notice high frame times on youtube videos.

I've seen stuttering many times in You tube videos.

In contrast, his gameplay was smooth and I feel it reinforces the importance of CPU to GPU balance that I mentioned in post #244.

As you can see from my slides above, even if you have high frame rate(fps) you may still have high frame times that will lead to stuttering.

Those examples you posted are 1.) different games and 2.) using HD7950 @ 1000 Mhz (which is way more powerful than his R7 240). With that mentioned, I do wonder how much larger he could have gone on the dGPU before running into stuttering problems.

 

[cbn]

I dont have a video for the Core i3 6300 but those two will do,

Win 10 64bit
Cat 15.11Beta

Ryse - 1080p High settings

BF4 MP 64player map - 1080p Custom 120fps Cap and Mantle

Huh. Wonder if I would notice that in a vid though . . . still interesting.

Also, a 6300 losing to an 8150 in average fps? lulz. I'd like to see a 7870k in that comparison. Or a Bristol Ridge when it comes out.

I dont know if you can notice that in a youtube video but i could definitely feel the difference while playing.

Battlefield 4 64 player had a minimum FPS of 1.

So if you made a you tube video of that gameplay I'm sure we would have noticed that.

 

[cbn]

Duplicate post.