AMD Ryzen 7 Performance: Windows 7 vs. Windows 10

So, essentially, Windows 10 needs some fixes applied to take care of the issue. Please, lets not make this into a stupid Linux vs Windows, Windows 10 vs Windows 7, or whatever vs thread, thanks. Oh, and thank you OP for posting this.
 
I knew Ryzen is harnessing power that is being hindered by Windows 10. Yeah it sucks for overclocking but a good 3.9-4ghz clock will tear through anything once Microsoft gets of their ass and fix Windows.
 
ManofGod said:
So, essentially, Windows 10 needs some fixes applied to take care of the issue. Please, lets not make this into a stupid Linux vs Windows, Windows 10 vs Windows 7, or whatever vs thread, thanks. Oh, and thank you OP for posting this.
Click to expand...

Yes and it is not my intention to start VS. discussion. I am testing this because I would like to know what is happening.
 
I cant tell in the counterstrike video, but in doom and tomb raider it appears the GPU's are clocked differently between windows 7 and windows 10. Ram usage also appears much higher on windows 7 for some reason. IM not familiar with the benchmark so im not sure if thats GPU ram utilized or system ram. Seems like a pretty big disparity though.
 
dgacioch said:
I cant tell in the counterstrike video, but in doom and tomb raider it appears the GPU's are clocked differently between windows 7 and windows 10.
Click to expand...

the clocks of the gpu depend on how the cpu can push it specially when not in the gpu limit
 
Last edited:
dgacioch said:
I cant tell in the counterstrike video, but in doom and tomb raider it appears the GPU's are clocked differently between windows 7 and windows 10. Ram usage also appears much higher on windows 7 for some reason. IM not familiar with the benchmark so im not sure if thats GPU ram utilized or system ram. Seems like a pretty big disparity though.
Click to expand...
7 will always show higher ram figures, that is the benefit of win 10 and WDDM 2.0.
 
BrianB said:
Click to expand...

Have you seen any difference with core, or rather thread usage? What I mean is do you have any games that are low thread count to compare against high thread count games? Although I do find it odd that 7 doesn't even try to use the SMT.
 
Rvenger said:
I knew Ryzen is harnessing power that is being hindered by Windows 10. Yeah it sucks for overclocking but a good 3.9-4ghz clock will tear through anything once Microsoft gets of their ass and fix Windows.
Click to expand...

Perhaps Microsoft has compiled their builds in such a way that they favor Intel CPUs? There weren't that many places putting a lot of effort into AMD FX CPU reviews or even the newer AMD APUs for that matter. This could be exposing a flaw that failed to be explored because of the seeming irrelevance of AMD CPUs for the last few years.
 
MrCaffeineX said:
Perhaps Microsoft has compiled their builds in such a way that they favor Intel CPUs? There weren't that many places putting a lot of effort into AMD FX CPU reviews or even the newer AMD APUs for that matter. This could be exposing a flaw that failed to be explored because of the seeming irrelevance of AMD CPUs for the last few years.
Click to expand...


Should we be accusing Microsoft of gimping AMD CPUs then? FX was an architectural mess, so I'd leave that conversation go.
 
Rvenger said:
Should we be accusing Microsoft of gimping AMD CPUs then? FX was an architectural mess, so I'd leave that conversation go.
Click to expand...

That wasn't my point. The reason why I asked is because it would make sense to focus optimization efforts for the dominant platform and in recent years that has unquestioningly been Intel. FX was bad, but may have been slightly worse as a result of a lack of effort to optimize for its architectural differences and with no one really paying attention to FX in reviews, there would not have been much said about it. Now that Ryzen has the potential to be competitive on a clock-for-clock basis with some of Intel's lineup, if there is a performance discrepancy due to that same lack of optimization effort, it might see some press coverage now.
 
BrianB said:
Yes and it is not my intention to start VS. discussion. I am testing this because I would like to know what is happening.
Click to expand...

Our continued thanks for these benchmarks. Really does make it realize how great and complete 7 really is.
 
Windows 10 sucks....yelling about this pile of crap OS for years now and noone believes me. I tested the crap out of Win10 vs. Win7 across shit load of setups AMD and Intel and Win 10 is just adware pile of crap OS. Do you think that scheduling works right on Xeon 14/28 vs Windows 7...no it does not. It doesn't affect as bad as Ryzen.
 
Luka said:
These video uploaded up there show different picture. This article didn't even bother showing real benchmark across different games.
Click to expand...

Yeah but it pushes his agenda of it being Ryzens fault and not a issue in Windows. Yet there are plenty of videos and forums posts that show otherwise.
 
Gideon said:
Yeah but it pushes his agenda of it being Ryzens fault and not a issue in Windows. Yet there are plenty of videos and forums posts that show otherwise.
Click to expand...

I have no agenda kid and I prefer windows 10.
 
Gideon said:
Yeah but it pushes his agenda of it being Ryzens fault and not a issue in Windows. Yet there are plenty of videos and forums posts that show otherwise.
Click to expand...

Multiple reviews have demonstrated that W10 is working fine, the last analysis from computerbase even recommends using W10 over W7 for RyZen. And AMD itself has officially stated that W10 works fine for RyZen:

We have investigated reports alleging incorrect thread scheduling on the AMD Ryzen™ processor. Based on our findings, AMD believes that the Windows® 10 thread scheduler is operating properly for “Zen,” and we do not presently believe there is an issue with the scheduler adversely utilizing the logical and physical configurations of the architecture.
Click to expand...

JustReason said:
he wasn't referring to you. He was referring to juanrga.
Click to expand...

In reality he was referring to anyone (including AMD) that doesn't say what he want to hear.
 
Monkey God said:
I am having a hard time seeing the data for all 13 games in that review.
Click to expand...

Same.

It looks like they average everything together to report nothing.

But there are specific games were Windows 7 does perform better.
 
Last edited:
JustReason said:
he wasn't referring to you. He was referring to juanrga.
Click to expand...

He hasn't given up yet? I blocked him as soon as I saw the name. I actually feel bad for him in some ways but , he made his own bed. Infamous in his own way.

Win 10 and WIn 7 just don't have the same kernel. The way they launch programs and utilize memory is not the same, therefore anything caching that memory is not behaving the same. Windows 10 is literally more guarded and executes more code to check what's running and where in the system those resource calls are headed. From VBS to to new ASLR implementations, everything Windows 10 does is different and adds some overhead in the name of security.

You might not like Win 10 because it takes away from DX 11 gaming or whatever, but I wouldn't trade the OS's personally. It's just time to move on imo. While I might "miss" my 75 Civic just because it was basic and economical doesn't mean I don't appreciate the air bags and crumple zones in my 016 Cherokee.

Sure they're different and it seems obvious some things were faster in Win 7. But when you get to the heart of it and think about why that is I'll give up those few per cent. The other complaints regarding Win10 privacy can be managed with very little work. Usually the privacy complaints come from people who hand over their lives to Google across all platforms without a thought.
 
Snowdog said:
Same.

It looks like they average everything together to report nothing.

But there are specific games were Windows 7 does perform better.
Click to expand...

Yep. Bottom line is if someone wants absolute best performance for specific games and applications, they should search on multiple reviews for those specific games and applications and avoid generalities.
 
Snowdog said:
It looks like they average everything together to report nothing.

But there are specific games were Windows 7 does perform better.
Click to expand...

And specific games like AoS where W10 does performs much better (23% faster).
 
So it's saying if you have Ryzen on Windows 10 you should be using High Performance power plan until the April patch.
 
Snowdog said:
Same.

It looks like they average everything together to report nothing.

But there are specific games were Windows 7 does perform better.
Click to expand...

Unfortunately every case of where it performs better in Win7 from what I have seen is when it is only using an Nvidia GPU, in fact the FPS performance was shown in one of the vids to be better in Win7 even though the SMT was not working right with Win7 (refused to use the 'SMT' related logical cores) while Win10 was using all logical cores and again was Nvidia GPU.
Looncraz did extensive testing with Win7 and Win10, he found fps did not improve for him while using AMD GPU.
If we can find other cases of Win7 and Ryzen with AMD GPU it would help.

Thanks
 
Also one needs a decent GPU to see any benefits with SMT, even on Intel and it is still very game dependent.
The problem with TechEpiphany vids linked is that they are using a 3GB GTX 1060 and this really is not enough grunt, and separately skewing the data as it seems Nvidia cards are the ones performing faster under Win7 rather than Win10 (need more data with AMD GPU testers).

Jump to 3min 22seconds to see just how the GPU affects SMT in terms of performance with the Intel 6700K, even their GTX1070 was not enough to see notable average fps improvements (minimum 1% framerates would probably stand out in a detailed analysis with the 1070) and needed higher GPU for avg fps increase.
But it still depends upon game tested.



Anyway it is another consideration, and why IMO it is better SMT is tested with specific multi-threaded applications and Benchmarks rather than games if the point and context is around proving SMT behaviour working and relative performance - remember games are incredibly challenging to design to scale well with many cores and even more so with SMT along with the GPU grunt required to truly see what is happening.
Cheers
 
Last edited:
Just to add and reiterate while context is applicable to Ryzen testing.
It is worth pointing out that the Hardware Unboxed video did not analyse the 1% and 0.1% minimum frames (would be perceived as stutter type effect), here the GTX 1070 would had seen improvements with Intel's 'Hyper threading' (SMT).
However most analysis including the OP vids on Ryzen + SMT are only looking at avg fps or current that changes rapidly on screen without it being detailed and without data captured-broken down, so really needs a decent GPU with a lot of grunt in these situations.

So a GTX1070 can still be good with the 6700K and SMT/Hyper threading at 1080p, just that it is more likely to improve the minimum 1% and 0.1% frametimes rather than a nice notable avg fps but it is all game dependent.
Cheers
 
The test needs an Intel CPU as a control group to see if Windows 10 is actually that bad.
 
realworld said:
The test needs an Intel CPU as a control group to see if Windows 10 is actually that bad.
Click to expand...

Intel's L3 cache on their CPUs is interconnected throughout all cores, isn't it? I might be falling behind on the conversation, but last I remember there was a hypothesis that the performance on AMD was more related to jumping across CCX clusters and moving information between the separate L3 caches. Has this been disproven? I also don't know if you could blame Windows 10, as much as even I may want to, for the disparity in SMT on/off scenarios because the implementation of SMT is different between the Intel and AMD architectures.

The Windows 7 performance really is interesting.
 
MrCaffeineX said:
Intel's L3 cache on their CPUs is interconnected throughout all cores, isn't it? I might be falling behind on the conversation, but last I remember there was a hypothesis that the performance on AMD was more related to jumping across CCX clusters and moving information between the separate L3 caches. Has this been disproven? I also don't know if you could blame Windows 10, as much as even I may want to, for the disparity in SMT on/off scenarios because the implementation of SMT is different between the Intel and AMD architectures.

The Windows 7 performance really is interesting.
Click to expand...

As demonstrated by reviews SMT-off only increases Ryzen average gaming performance by 1% and 3%. As demonstrated by reviews 4+0 is on average only 5% faster 2+2. As demonstrated by reviews Ryzen is 2% faster on W10 than W7. Meanwhile AMD has given official statement that (i) W10 works fine, (ii) there is no scheduler issue, and (iii) there is no SMT issue.
 
juanrga said:
As demonstrated by reviews SMT-off only increases Ryzen average gaming performance by 1% and 3%. As demonstrated by reviews 4+0 is on average only 5% faster 2+2. As demonstrated by reviews Ryzen is 2% faster on W10 than W7. Meanwhile AMD has given official statement that (i) W10 works fine, (ii) there is no scheduler issue, and (iii) there is no SMT issue.
Click to expand...

So best case scenario we are talking about 10% performance difference between ideal circumstances and normal?
 
MrCaffeineX said:
So best case scenario we are talking about 10% performance difference between ideal circumstances and normal?
Click to expand...


Yes, and add another 10% for high speed ram on top of that and you have a significant jump in performance vs. a lot of the release reviews.
 
You must log in or register to reply here.
Back
Top