• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Ryzen and Gaming results.

I'm late to this thread so it may have already been said but isn't the IPC actually really good? What AMD do need is equal clocks.

People keep saying the R5 should be better but they've already said they are going to be 3.7ghz so how is that going to be faster than a R7 at 3.7ghz???
 
Overlag said:
I'm late to this thread so it may have already been said but isn't the IPC actually really good? What AMD do need is equal clocks.

People keep saying the R5 should be better but they've already said they are going to be 3.7ghz so how is that going to be faster than a R7 at 3.7ghz???
Click to expand...

IPC wise Ryzen is around 6-8% behind Kabylake. The issue is their process, 14nm LPP doesn't like very highclocks; it gets extremely hot when clocked up due to needing more voltage.
As a result R5 clocking higher is rather unlikely, but could happen due to less cores, and the 4Core having a single CCX vs 2. Other than that, clock for clock they should be near identical; with the R5 pulling ahead on some games because it'll have less latency due to not having two CCXs; and instead 1one.

Even so Ryzen does amazingly well in everything; with it's only pitfall being gaming performance at the moment.

Check out the Game Benchmarks video on the previous page, and the chaps other videos. Fantastic comparisons running raw footage with stats overlays.
 
I imagine performance will get better over time as more developers adopt DX12/Vulkan and learn to implement it properly. Right now it seems anything above four cores is wasted compute power.
 
BIOS on the MSI Tomahawk motherboard has an option to shut off cores. So in theory if R5 chips are R7 rejects then that can be tested already.
 
N19h7m4r3 said:
Even so Ryzen does amazingly well in everything; with it's only pitfall being gaming performance at the moment.

Check out the Game Benchmarks video on the previous page, and the chaps other videos. Fantastic comparisons running raw footage with stats overlays.
Click to expand...

See now that's the problem... calling it a "pitfall", totally uncalled for, and exactly what Intel would want people to believe. It's a harsh word that paints a very negative picture, despite the fact that in some games Ryzen is level pegging, while in a few it even has the edge. Yes it loses in a fair few, but how much of that is down to lack of optimisation, motherboard BIOS, RAM etc.? An unknown quotient of course, but there are clearly lots of factors at play here, and in time it seems inevitable the gap will close once these areas are improved upon and the issue ironed out (which they will be). Lots of people are also not seeing the 'minimum' FPS is often higher with Ryzen than Kaby, which is significant, and something that will be far more noticeable than maximum, especially when those maximums are so high.

I know you did say "at the moment", but it's wrong to call gaming a Ryzen pitfall, when it's more of a bump in the road.
 
mmj_uk said:
DX12/Vulcan are meant to make games even less CPU limited so if anything anything above four cores will be even more wasted.
Click to expand...

It is supposed to be less taxing on the cpu compared with DX11 but I would imagine there is also a way to make use of that freed up compute power and translate that in to performance otherwise it make the need for the latest and greatest cpus in the future null and void in regards to gaming at least. Which surely cannot be the case.
 
Overlag said:
I dunno it also is supposed to make it easier to use more threads? (in the long run?)
Click to expand...

It is but at the same it's reducing the overall overhead which helps lesser core processors too, I think when Vulcan benchmarks first came out it helped boost AMD's quad core APU's the most.
 
mmj_uk said:
It is but at the same it's reducing the overall overhead which helps lesser core processors too, I think when Vulcan benchmarks first came out it helped boost AMD's quad core APU's the most.
Click to expand...
Overhead and threads are 2 different things, you can have high over head but the game will only use one thread, you can have low overhead but the game is highly threaded, low over head is not going to make a game that likes lots of threads
work just as well on CPUs that have less cores.
 
mmj_uk said:
DX12/Vulcan are meant to make games even less CPU limited so if anything anything above four cores will be even more wasted.
Click to expand...

That's a very slightly off understanding of the advantage of DX12/Vulcan you're referring to. And leads to a wrong conclusion. Yes, one of the advantages of DX12 and Vulcan is to try and make the CPU less of a bottleneck, but not by just lowering the CPU requirements in some abstract way. It helps reduce the bottleneck in a very specific and describable way which is to enable multiple threads to feed the GPU at the same time. It doesn't speed up the CPU in some way that makes the CPU less of a bottleneck, it broadens it out so greater use can be made of multiple threads. Seen in detail, it then becomes obvious that rather than DX12 and Vulcan making more cores (sorry, moar corez) less necessary, is better thought as making higher core / thread counts more useful / easier to take advantage of.

Additionally, you're essentially arguing against the need to upgrade at all. That, for example, an FX-8350 is a viable gaming processor. Which it is. But the general area of discussion is between AMD and Intel current equivalents, not between current chips and whether people actually want the high-end or not. Saying: "you don't need to upgrade" may be true for any given gamer and / or budget, but is not addressing whether to look at Intel or AMD if you are which is where I understand the debate to be.
 
muon said:
BIOS on the MSI Tomahawk motherboard has an option to shut off cores. So in theory if R5 chips are R7 rejects then that can be tested already.
Click to expand...

Someone already has in the main thread but it was lost in the pages. I think it was something like a 200mhz increase in overclock with 4 cores disabled. Couldn't be sure though as I can't find the post now.
 
AMD should have had a bigger memory bus. Like the X99 chips. Seems to improve a lot if you're u can give it more memory bandwidth
 
Overlag said:
AMD should have had a bigger memory bus. Like the X99 chips. Seems to improve a lot if you're u can give it more memory bandwidth
Click to expand...

Maybe it was held back for cost and/or so they can show larger improvements for Zen+?

Slightly cynical, but it's probably better for their image and investors if they pretty much matched Intel at launch, then made nice improvements on Zen+, than if launch Zen was at the bleeding edge already and they could barely improve it moving forward.
 
dual is just as fast so probably thought lower costs and keep same performance gaming wise.than going quad.cheaper mobos cheaper full stop.
 
Has anyone tried windows 10 game mode yet?

http://www.theverge.com/2017/1/27/1...e-mode-now-available-insiders-preview-program

Microsoft gave details on the upcoming Windows 10 Game Mode feature earlier this week, and today began to roll it out to users in the Windows 10 Insider Preview program as part of the latest 15019 preview build.

If you’re trying out the build, Microsoft is warning users that there could be a few bugs. In a statement released to The Verge, the company noted that the latest build includes “a few platform related bugs, unrelated to the new features included within the build, that are impacting the ability to play some popular games,” which is something to be aware of if you’re looking to try out Game Mode.

Windows 10 Game Mode allows users to get improved performance for gaming from their hardware by prioritizing CPU and GPU resources to optimize frame rate. The feature will be included as part of the Windows 10 Creators Update that’s coming later this year.
 
Razor Time said:
Not sure if this has already been posted but its an interesting video as to why gaming performance is different to productivity.

Click to expand...
That's pretty interesting. So latency between cores on different CCXs is 3.5x as high as between cores on the same CCX. Some rough numbers from their graphs:

Intel SMT latency (same physical core, different logical core): 15 ns
AMD SMT latency (same physical core, different logical core): 25 ns

Intel core latency (different physical core): 80 ns
AMD core latency (different physical core, same CCX): 45 ns
AMD core latency (different physical core, different CCX): 145 ns

It also shows that Windows does understand AMD's SMT implementation, at least in terms of the core layout. Patching Windows to understand the CCX implementation would very likely garner the biggest improvement.
 
I would expect CCX to CCX transfers to be one of the big areas of effort for improvement coming to Zen+/Zen++
 
You must log in or register to reply here.
Back
Top Bottom