Ryzen with 3600MHz RAM Benchmark

dcds1 said:
So you then agree that people need to get out their high horse and accept that everyone plays with other resolutions and graphic cards. I am sure nVidia has sold a whole bunch more 1070 graphic cards than 1080 so that would make more of a real world experience than a 1080 graphics card.
Click to expand...


Yes and no. To test a CPU, you want to eliminate any potential bottleneck, so testing with a 1070 is not the worst choice, but it is not the best TO TEST THE CPU. people are 3x more likely to upgrade their video card over their CPU because video cards progress in performance per dollar MUCH faster than CPUs. In 18 months, the 1070 will be considered an "entry level mid-range card" just as the 970 is now. And god forbid anyone try and game with a 770. The 770 is only a 4 year old card. The person who bought a 770 four years ago probably upgraded their card to a 1070 recently. If they bought an AMD 8350 because it was close enough to the i5 3570 in 2013's "real world gaming tests" how much do you think they would be kicking themselves now?

THAT is why eliminating the GPU bottleneck is important.
 
Junkboy said:
Currently very few boards have the bios capability to mess with the blck so only a handful of users can even test it. And with all the bricks Asus caused with their friendly windows auto bios updater even less boards for folks to test with.

My poormans gaming 3 doesn't but I have some F4-3600C16D-16GTZ on the way so I can be ready if and when bios get support for higher speeds.
Click to expand...

Sweet! Good luck hopefully you can even hit those speeds. I am interesting in Ryzen quite a bit. But I think I might wait for the bugs to get kinked out. Also that X399 chipset is coming soon too!
 
CSI_PC said:
You are dismissing it too quickly.
Check the 3200MHz Ryzen against both the vid and Eurogamer result I provided and then look at the OC.
It is clear there is a bottleneck being caused by the GTX1070 and this is not as noticable on Ryzen as the CPU bottleneck is right around the limit of the GTX1070.

What I showed is not an outlier, because nearly every result by various reviews involving tests using at least a GTX1080 has the Intel 7700K performing much better than a Ryzen shown in the video - context here is absolute CPU performance rather than gaming reality with lower GPUs.
The outlier is the video test using a GTX1070 to prove absolute CPU performance because it skews the results.

There is a reason why Eurogamer/Digital Foundry went with a Titan Pascal to test at 1080p when looking at absolute CPU performance, and you should put more weight on their testing methodology and results than the youtube vid for reasons that are pretty clear when comparing trend results.

Cheers
Click to expand...
Absolute CPU performance? Most games do not stress 8 core cpus. Anyways how come, if RyZen is some how deficit in gaming that a 1700x OC to 3.9ghz beat the pants off of a 7700K in Dues Ex Mankind Divided using DX 12?
http://www.tomshardware.com/reviews/amd-ryzen-7-1700x-review,4987-4.html

Are not games of the future will be more like Dues Ex Mankind Divided? Also in Division DX 12 RyZen was very close. Another title more future like.

The game benchmarks at low resolutions (yes 1080p I consider low resolution) does not equate to cpu benchmarks for non-gaming where it slaughters in many cases the 7700K. So what use are they? Gaming benchmarks at 1080p for current games most likely will not reflect future gaming ability. The winner over time between the 8 core RyZen and even the 6 core RyZen and 7700K is most likely RyZen due to the more cores that can come into play for games. Once games start to be more RyZen aware and programmed for, where as the 7700K already has that, RyZen I do believe is the much better gaming cpu in the long run, 6 core and up.

For current generation hardware I see no significant game play experience differences for Intels best gaming cpu and RyZen anyways.
 
noko said:
Anyways how come, if RyZen is some how deficit in gaming that a 1700x OC to 3.9ghz beat the pants off of a 7700K in Dues Ex Mankind Divided using DX 12?
Click to expand...
I'm not sure I would call a 1fps lead "beat the pants off", but it's a win nonetheless.
 
noko said:
Absolute CPU performance? Most games do not stress 8 core cpus. Anyways how come, if RyZen is some how deficit in gaming that a 1700x OC to 3.9ghz beat the pants off of a 7700K in Dues Ex Mankind Divided using DX 12?
Click to expand...
I am not sure you can call suspect measurement (suspect enough tom's proceeded to ask Eidos about it) a "beat the pants off". I mean, hell, fx-8350 sits within 1 fps away from 6900k. Meaning either it is a GPU benchmark in which case Ryzen result is suspect. Or it's worse than that.
noko said:
Are not games of the future will be more like Dues Ex Mankind Divided?
Click to expand...
AMD assisted? Maybe.
noko said:
For current generation hardware I see no significant game play experience differences for Intels best gaming cpu and RyZen anyways.
Click to expand...
For most AAA games that is indeed the case. Except then your claims about Ryzen being better gaming CPU in long-run is even larger pile of BS than claims of people claiming Ryzen sucks because it has twice the memory latency.
 
lolfail9001 said:
For most AAA games that is indeed the case. Except then your claims about Ryzen being better gaming CPU in long-run is even larger pile of BS than claims of people claiming Ryzen sucks because it has twice the memory latency.
Click to expand...

For the time being yes... Good progress has already been made on tweeking the ryzen platform. I'm not sure why we can't all agree that both intel and AMD are offering good CPU products.
 
noko said:
Absolute CPU performance? Most games do not stress 8 core cpus. Anyways how come, if RyZen is some how deficit in gaming that a 1700x OC to 3.9ghz beat the pants off of a 7700K in Dues Ex Mankind Divided using DX 12?
http://www.tomshardware.com/reviews/amd-ryzen-7-1700x-review,4987-4.html

Are not games of the future will be more like Dues Ex Mankind Divided? Also in Division DX 12 RyZen was very close. Another title more future like.

The game benchmarks at low resolutions (yes 1080p I consider low resolution) does not equate to cpu benchmarks for non-gaming where it slaughters in many cases the 7700K. So what use are they? Gaming benchmarks at 1080p for current games most likely will not reflect future gaming ability. The winner over time between the 8 core RyZen and even the 6 core RyZen and 7700K is most likely RyZen due to the more cores that can come into play for games. Once games start to be more RyZen aware and programmed for, where as the 7700K already has that, RyZen I do believe is the much better gaming cpu in the long run, 6 core and up.

For current generation hardware I see no significant game play experience differences for Intels best gaming cpu and RyZen anyways.
Click to expand...
What seems to get glazed over is that some, well a few, reviews actually did those low resolution tests and we know that outcome, but then did a REAL gameplay at 1080p and there was no difference at all in most games. This 1070 debate is lacking in that the 1070 was benched using real world gameplay not CPU benching as it were and to be honest it is the true result that should have more impact on a purchasers decision. Hell the Ryzens when used under real world conditions performed better than all Intels, albeit only a fps or 2, mostly in minimums as the avg were all equal.
 
noko said:
Absolute CPU performance? Most games do not stress 8 core cpus. Anyways how come, if RyZen is some how deficit in gaming that a 1700x OC to 3.9ghz beat the pants off of a 7700K in Dues Ex Mankind Divided using DX 12?
http://www.tomshardware.com/reviews/amd-ryzen-7-1700x-review,4987-4.html

Are not games of the future will be more like Dues Ex Mankind Divided? Also in Division DX 12 RyZen was very close. Another title more future like.

The game benchmarks at low resolutions (yes 1080p I consider low resolution) does not equate to cpu benchmarks for non-gaming where it slaughters in many cases the 7700K. So what use are they? Gaming benchmarks at 1080p for current games most likely will not reflect future gaming ability. The winner over time between the 8 core RyZen and even the 6 core RyZen and 7700K is most likely RyZen due to the more cores that can come into play for games. Once games start to be more RyZen aware and programmed for, where as the 7700K already has that, RyZen I do believe is the much better gaming cpu in the long run, 6 core and up.

For current generation hardware I see no significant game play experience differences for Intels best gaming cpu and RyZen anyways.
Click to expand...

Totally wrong context to all my posts here.
Absolute performance from perspective of bottlenecks :)
You cannot see the absolute performance of a CPU in a game even if it does not stress CPU cores when the GPU is bottlenecked; in fact the GTX1080 shows the CPU limitation even in games with your view that do not stress cores because Ryzen CPUs bottleneck before the 7700K.
Come on my point and context and with actual facts has been quite clear on this and now your skewing it, along with Justreason.

Even you must notice the trend difference and gap between a GTX1070 and GTX1080/1080ti with the gap in games in the comparison between Ryzen and a 7700K.
But now because you done that post we are going in a circle even though I explained the difference of absolute performance context for games and gaming reality with what is used but consideration needs to also be for future upgrades and how soon hit the CPU bottleneck (will affect some and not others as everyone upgrades different times and with different tier of GPUs).

Both points are valid and have a place in this conversation, but as I keep saying the absolute performance scope (for gaming as this is what the thread is about) should not be bottlenecked by the GTX1070, and a consideration influencing both points is how long before owners upgrade to a Volta 2070 while looking to keep the CPU for another 2-3 years.
That is a very nuanced debate as it would mean the bottleneck with current modern games would again be the Ryzen CPU without gains beyond a GTX1070 but it could also be seen future next gen games may bottleneck Volta 2070 GPU even with that power at 1080/1440p, along with the point it is cheaper to upgrade AMD CPUs than Intel and maybe better game design will remove some of that Ryzen bottleneck.

Cheers
 
Last edited:
Could just use some emulator to never have to worry about GPU performance impacting the end result. :p I remember how the fps boost you got from overclocking the CPU was linear with the amount it had been overclocked.
 
ZeroBarrier said:
Wait, are you seriously surprised to see Ryzen beat an i7 7700K in a biased and flawed review?

I mean, we can probably get Piledriver to beat an i7 7700K with enough bias. So why be surprised to see Ryzen beat it in a biased review...
Click to expand...

Why is this review flawed? Had 3200Mhz ryzen and i7 in it what the hell else is the issue now.
 
shad0w4life said:
Why is this review flawed? Had 3200Mhz ryzen and i7 in it what the hell else is the issue now.
Click to expand...


Technically it is not. Its what the reviewer had with resources on hand. He had a 5ghz Kabylake with a 2200mhz GTX 1070 running 1080p. Last time I checked the 1070 was a 1440p card and the usual whiners are crying GPU bottleneck. The 1070 was probably running within 10% of a 1080 at those clocks. If you read his former reviews he explains why his system is the way it is. It was a good review and it shows that Ryzen has enough chops when RAM speed is higher.

Ryzen is selling well because they have more chips than motherboards right now. Which I am surprised that how many chips AMD produced. Most are still waiting on the motherboards which are in a way hindering the possible chip sales.
 
lolfail9001 said:
It's not, if you are going for ultra settings. And even at 1080p high, reviewer was running, Titan XP pulls out much higher fps. Ergo, 1070 is the bottleneck.
Click to expand...


Need someone to repeat this with a higher end card on the 7700k.



bench.png


Capture.PNG
 
Last edited:
Rvenger said:
Need someone to repeat this with a higher end card on the 7700k.



View attachment 20493


View attachment 20494
Click to expand...

To re-iterate:

CSI_PC said:
OK I went back through the vid to see if I could compare any aspect to another review highlighting the point that the GTX1070 is bottlenecking the 7700K for the discussion and context around absolute CPU performance.
One stands out as it can be compared with Eurogamer/Digital Foundry review where they compared the 7700K to the 1800X with RoTR, the 1800X had 3200MHz memory while the 7700K had 3000MHz memory.

Youtube vid with GTX1070:
1700X at 3.97GHz/3600MHz: 87.1 fps
1700X at 3.97GHz/3200MHz: 86.9 fps
7700K 5GHz/3200MHz: 94.3 fps

Eurogamer/Digital Foundry with Titan Pascal:
1800X no OC/3200MHz: 85.8 fps
7700K no OC/3000MHz: 126.5 fps

Both at 1080p.
The 7700K is being bottlenecked by the GTX1070 while the 1700X/1800X for now (will improve but unfortunately none of us know by how much and how long it will take) are the bottlenecks in their setup.
Obviously not every game will be so, but quite a lot will be limited by the GTX1070 in the context of testing the limits of a 7700K at 1080p.
Cheers
Click to expand...


Also GamersNexus used a custom GTX1080 and here are their results:

ryzen-r7-1800x-gtav.png




Unfortunately the youtube vid is bottlenecking Intel with his settings/resolution even with the GTX1070 while Ryzen is just around the limit of said GPU.
Even the custom GTX1080 is bottlenecking the Intel CPUs with those settings, anyway it is clear with Gamers Nexus results the Ryzen CPUs are bottlenecking the GTX1080.
Cheers
 
Last edited:
Wasn't that a release review? Didn't we already have BIOS update and windows patches that solved some of the issues already. I think so. Honestly, I am going to resort to my own testing since I have this gear in house for the most part. I will come to my own conclusion instead of you guys regurgitating launch reviews. I game at 4k and also don't use my system just for gaming either. So I could care less, again. If I want superior single threaded performance on some of the games that need it, I have my Skylake setup sitting next to me for that. We can revisit these performance issues once time passes some more because there's just too much inconsistent data out there, everywhere.
 
Rvenger said:
Wasn't that a release review? Didn't we already have BIOS update and windows patches that solved some of the issues already. I think so. Honestly, I am going to resort to my own testing since I have this gear in house for the most part. I will come to my own conclusion instead of you guys regurgitating launch reviews. I game at 4k and also don't use my system just for gaming either. So I could care less, again. If I want superior single threaded performance on some of the games that need it, I have my Skylake setup sitting next to me for that. We can revisit these performance issues once time passes some more because there's just too much inconsistent data out there, everywhere.
Click to expand...
You are not going to get over such a bottleneck without certain changes that would include.
- Changing the thread behaviour/'optimising' as much as possible for Ryzen with each game.
- Changes to the Ryzen; whether architecture or firmware that may improve gaming performance behaviour through the Data Fabric/cache coherency/etc.
- Microsoft's Gaming Mode as it controls threads-to-cores in a different way while making distinction between them and more general/OS related ones, in this context does not really make much of a difference for Intel but would help Ryzen's challenges. This is a broader solution than doing it on a game dev level.

And for these it is very difficult to tell just how much improvement it would make to said bottleneck behaviour against the more powerful GPUs such as GTX1080 and higher.

But that is going beyond context of this thread and your post about comparing bottlenecked performance of the youtube vid to that of Eurogamer/Gamers Nexus/etc that show that bottleneck happening.
Time will tell and we really do need at least one of these updates I am talking about before we can even work out what gains we may see for Ryzen.
But this really only applies for those wanting to game at 1080P with something better than a GTX1070 for now or considering possibly how it may impact them if looking to buy a Volta 2070 in 2 years time but fingers crossed any of those updates mentioned helps.
Cheers
 
Last edited:
Rvenger said:
I game at 4k and also don't use my system just for gaming either.
Click to expand...
I've had my Ryzen rig since launch (see sig). 4K gaming performance is great. For example, I can get around 65 - 70 fps in GTA V (4K everything maxed except advanced options). And this is with RX 480 CF, impressive for around $450 of GPU. Some other games aren't multi-GPU optimized, so I'm playing a few older titles now, but it's still maintaining at 4K.

Also, I briefly had a 1080 Ti in this rig and compared it to my other 4K rig (not in sig, but using a 4790K + TitanXP) and performance was just about equal (if not a hair better on 1800X + 1080 Ti). With the i7/TXP I got around 6600 in Firestrike Ultra but around 6700 with Ryzen/1080Ti. In gameplay they were about equal to my eye.
 
Last edited:
So what ram will run @ 3600 mhz.
I updated to Asus latest bios and now my ram wont post past 2666 no matter what I throw at it.

I have CHVI motherboard. I am really curious as to why this happened?

Anyone care to add their ram settings from successful ram overclocking on their Crosshair VI? I am using DDR4-3200 rated QVL listed Corsair LPX. THinking of trading it back to Microcenter for even higher rated RAM and taking a gamble at it posting high clocks. Was thinking of the new LED series as I love the look of the lights on the new Corsair ram.

And with these benchmarks sigh... when are we going to accept that FPS past 90fps in any title at max settings is all a moot point. Sure a CPU can deliver more FPS overall but in the big picture there is no discernable and USEFUL difference that can be detected to the human senses. Truly I have 120FPS eyes that seem to not tell the difference past 90fps. Its biology and no human being can actually perceive, truly percieve, more than a certain limit of FPS, I can't remember exacly where but it certainly is not 179 fps lol like Mr. 5.1ghz 7700 is doing. I just get tired of people saying the 7700K is absolutely superior to the Ryzen line becaues of just gaming. I think many people who build a PC for non-gaming with gaming on the side would argue that there is no legitimate argument to the claims of all these anti-amdr's are making daily. Even down to earth Intel junkies will admit the Zen is a multitasking monster, sure maybe not the top gaming chip, but down to Earth Intel guys also do NOT use their PC only for gaming.

What I am finding interesting with my Zen is that performance scales quite considerably the higher I can clock my ram. That was until somereason it stopped clocking past 2666mhz. I am confused beyond measure as to why. I am tempted to waterblock the ram and voltage pump the piss out of it to see what it can do. It could also be the memory controller needs a voltage bump on the CPU but I wont do this until my waterblock adaptor arrives from Performance PCS this week and I can get my raystorm connected.
 
Last edited:
itt: a bunch of people unable to comprehend that 3200mhz ram with tight timings can perform just the same as 3600mhz ram with loose timings.




And like many others, I wish there was a list of motherboards that currently support 3000mhz+ ram. I really want to uprade to Ryzen, but there aren't any good motherboards to buy yet. Maybe in another month or two BIOs fixes will finally address the RAM issues.
 
Odio said:
itt: a bunch of people unable to comprehend that 3200mhz ram with tight timings can perform just the same as 3600mhz ram with loose timings.




And like many others, I wish there was a list of motherboards that currently support 3000mhz+ ram. I really want to uprade to Ryzen, but there aren't any good motherboards to buy yet. Maybe in another month or two BIOs fixes will finally address the RAM issues.
Click to expand...
it is just a tad more complicated than that. The infinity Fabric in Ryzen is tied to ram speed, ergo why everyone one is shooting for higher. I only expected to hit 1866 on my 8350, but Got 2400 and was impressed and quite surprised. I get the race to "how high".
 
Odio said:
itt: a bunch of people unable to comprehend that 3200mhz ram with tight timings can perform just the same as 3600mhz ram with loose timings.




And like many others, I wish there was a list of motherboards that currently support 3000mhz+ ram. I really want to uprade to Ryzen, but there aren't any good motherboards to buy yet. Maybe in another month or two BIOs fixes will finally address the RAM issues.
Click to expand...

If you watched the video, it looks like ram timing do not matter when it comes to Ryzen. Speed of the memory makes a difference, not ram timing.

Of course that is what the video says. I would like to see other people test the results. He claims the infinity fabric rely's heavily on memory speed not timings.

Also why are people complaining about the video card used? Why don't we talk about memory speed which is what the video is about. Should get back on topic.
 
Rvenger said:
Wasn't that a release review? Didn't we already have BIOS update and windows patches that solved some of the issues already. I think so.
Click to expand...
No update so far affected actual performance of Ryzen. Nor will it. It may make motherboards less of a pain than they are, but they are not going to pull improvements out of thin air.
Rvenger said:
I game at 4k and also don't use my system just for gaming either. So I could care less, again.
Click to expand...
Valid, though it begs the question of why do you even bother wasting your time arguing about it.
 
Brackle said:
it looks like ram timing do not matter when it comes to Ryzen. Speed of the memory makes a difference, not ram timing.

Of course that is what the video says. I would like to see other people test the results. He claims the infinity fabric rely's heavily on memory speed not timings.
Click to expand...
There is a mounting body of evidence to support the hypothesis.

 
Romeomium said:
Currently running @ 2933 on a single stick, but I may be limited by the gigabyte board. I have a thread under Memory & Mobos where I am updating progress. It's a micron stick, as I could not locate Samsung B-Die ECC UDIMMs. Still trying though!
Click to expand...
Thanks for the compilation of ECC UDIMM info thus far in your thread!
 
Looking like a much more optimized Ryzen build will be destroying Intel across the boards in six months or so.

So where will Intel be when Zen+ releases in 2018 already fully optimized?
 
joyzbuzz said:
Looking like a much more optimized Ryzen build will be destroying Intel across the boards in six months or so.

So where will Intel be when Zen+ releases in 2018 already fully optimized?
Click to expand...
Selling their Kaby lake 4 core X299 processors using quad channel DDR 4 ram at 3000+ speeds.
 
joyzbuzz said:
Looking like a much more optimized Ryzen build will be destroying Intel across the boards in six months or so.

So where will Intel be when Zen+ releases in 2018 already fully optimized?
Click to expand...

Very weak troll bait. Even ignoring that CPUs are only part of Intel's business, Ryzen isn't destroying anything of Intel's thus far. It's a decent competitor for the price; nothing more and nothing less.

noko said:
Selling their Kaby lake 4 core X299 processors using quad channel DDR 4 ram at 3000+ speeds.
Click to expand...

We've known Intel's roadmap for quite some time now. I believe his post was meant fasciously as troll bait.
 
CSI_PC said:
To re-iterate:




Also GamersNexus used a custom GTX1080 and here are their results:

ryzen-r7-1800x-gtav.png




Unfortunately the youtube vid is bottlenecking Intel with his settings/resolution even with the GTX1070 while Ryzen is just around the limit of said GPU.
Even the custom GTX1080 is bottlenecking the Intel CPUs with those settings, anyway it is clear with Gamers Nexus results the Ryzen CPUs are bottlenecking the GTX1080.
Cheers
Click to expand...
GTA V is not a future type coded game - yesterday more Intel optimized game. I have no problem with my 3440-1440p maxing out a constant 60fps in this title. No cpu, graphics card at this time can improve my game play because I am VIDEO MONITOR LIMITED. Which is a new concept for many I take. My hardware is driving my monitor at it's fastest speed with room to spare. My eX Fx 9590 could not do that.

Looking at DX 12 newer titles the differences look to become much smaller where for some reason Dues Ex DX 12 for RyZen smokes Intel. Maybe a testing error but I do find interesting. The benchmark below indicates a limitation on both platform but in this case never explored. Neither cpu speed differences made that much overall in the overall benchmark. So what is going on here, why could RyZen process the data faster than any Intel CPU? Is the larger cache of RyZen the reason? Can other titles take advantage of that? I would think so if the case.

Bench.jpg

Now with all of these low resolution gaming bench that is suppose to be testing the cpu - where is the corresponding CPU usage for a given benchmark? Would running a benchmark that the cpu is only using 30% or less capability would even be considered a cpu test? I just don't see it that way. It does hint at other limitations which seems like are never explored but some nonsensical general conclusion such as RyZen sucks at gaming or not best at gaming BS without even trying to dig a little deeper.

Now the ram speed was at DDR 2666, which is the max rated speed for single sided Dim, 2 modules. I am sure the results would be better at DDR 4 3200 but not everyone will be able to achieve that even with the best sticks of ram. Some RyZen IMC's will just not do it.
 
ZeroBarrier said:
Very weak troll bait. Even ignoring that CPUs are only part of Intel's business, Ryzen isn't destroying anything of Intel's thus far. It's a decent competitor for the price; nothing more and nothing less.



We've known Intel's roadmap for quite some time now. I believe his post was meant fasciously as troll bait.
Click to expand...
Maybe maybe not. I am sure Intel will do just fine. Intel's price may need to be adjusted but I am sure when the X299 launches it will be in a better state or should I say ready to launch then RyZen was. RyZen just made everything way more interesting and it is a very strong product for the price. Minus some rather large headaches some folks are having building a new system.
 
FYI there are other boards not just the CH6. So there really isn't an excuse for "I cannot get a board for my Ryzen".
BTW I have 2 x370 Boards.
I am currently playing with my 1700x on the Biostar Gt7. I also have ASUS Prime.
I like the Biostar better because of the connections.
I was kinda upset cause I had to RMA the First Biostar and orded the Prime to replace it. And they sent me another Biostar when I thought they would refund.
Memory matters more on the Ryzen platform. Still cannot get the Geil 2400 DDr4 to boot on either board. Gskill 2167 fury with 14 cas timings works fine! Better yet on the Biostar if I loosen the timings, on a trident cas 14, 3200 to 16-16-16-36 it boots at 3200. CAS 14 @2933 works on both boards.14-14-14-34
Since I have to buy 4 cards I could only afford the 1060, I am so ashamed.:woot:
FWIW I want you to explain to my wife and kids "Daddy really needed the 1080" so you all get the 1050Ti. That's really gonna go over well.:eggface:
No I am not going to test the prime as this will be going with an 1600x on the wife system or maybe 1 of the kids.

For what I do this 1700x is really fine.
 
syyDOFFK.jpg


RAM does help a significant amount, that is trite and was well known months before release. Interestingly Aida is also showing a Ryzen evolution

Per a member of AT

DDR4 2133 / 1000mhz

Aiada64_CacheMem_3.8GHz.png


After

DDR4 2600 / 1300mhz

AIDA_Cachmem_03272017.png


Just making a jump from entry level DDR4 to 2600 raised bandwidth 10%, dropped latency on memory over 20% and L2 and L3 latency went down significantly, it is again supporting the notion that once DDR4 3400 becomes status quo for Ryzen the performance will dramatically improve. All of this was achieved on ucode
 
I'm pretty set now on getting ryzen 1600, anyone know the specific ram/mb combos that are more likely to be able to support 3200 - 3600 MHz ram?
 
tybert7 said:
I'm pretty set now on getting ryzen 1600, anyone know the specific ram/mb combos that are more likely to be able to support 3200 - 3600 MHz ram?
Click to expand...

That is constantly changing, some vendors have working 3400 now but there is still much development, I know Asus and MSI have B350 and X370 up to 3400/3200 stable and BCLK tweaking.

I quite like the Gigabyte Gamer 3 with a nice kit of 3200mhz RAM should give you solid performance.
 
View attachment 20494
[/QUOTE]
noko said:
GTA V is not a future type coded game - yesterday more Intel optimized game. I have no problem with my 3440-1440p maxing out a constant 60fps in this title. No cpu, graphics card at this time can improve my game play because I am VIDEO MONITOR LIMITED. Which is a new concept for many I take. My hardware is driving my monitor at it's fastest speed with room to spare. My eX Fx 9590 could not do that.

Looking at DX 12 newer titles the differences look to become much smaller where for some reason Dues Ex DX 12 for RyZen smokes Intel. Maybe a testing error but I do find interesting. The benchmark below indicates a limitation on both platform but in this case never explored. Neither cpu speed differences made that much overall in the overall benchmark. So what is going on here, why could RyZen process the data faster than any Intel CPU? Is the larger cache of RyZen the reason? Can other titles take advantage of that? I would think so if the case.

View attachment 20548

Now with all of these low resolution gaming bench that is suppose to be testing the cpu - where is the corresponding CPU usage for a given benchmark? Would running a benchmark that the cpu is only using 30% or less capability would even be considered a cpu test? I just don't see it that way. It does hint at other limitations which seems like are never explored but some nonsensical general conclusion such as RyZen sucks at gaming or not best at gaming BS without even trying to dig a little deeper.

Now the ram speed was at DDR 2666, which is the max rated speed for single sided Dim, 2 modules. I am sure the results would be better at DDR 4 3200 but not everyone will be able to achieve that even with the best sticks of ram. Some RyZen IMC's will just not do it.
Click to expand...


Why are you saying this to me when it was Rvenger who did the request to compare against the Youtube Vid and I provided what he asked for!!!!
And Deus Ex Mankind is NOT a next gen title it is rather fubar tbh, in fact Tom's publication has been trying to get answers from the dev on why it is behaving that way; you do not think it strange that 7600k/7700k/6900k all have the same fps within 0.2 fps when they are left to their default core frequency and boost settings?

With regards to memory, I have linked data also showing improvements of memory frequency works equally for both Intel and Ryzen in games as Eurogamer has done this test on both and I posted the data, it is not unique to Ryzen.
Here it is again the request regarding GTA V:

Rvenger said:
Need someone to repeat this with a higher end card on the 7700k.



View attachment 20493


View attachment 20494
Click to expand...

Cheers

Edit:
And unfortunately seems Paul at Tom's used internal benchmark of Deus Ex Mankind Divided, so that really does not help and would raise into questions any result.
 
Last edited:
CSI_PC said:
View attachment 20494
Click to expand...



Why are you saying this to me when it was Rvenger who did the request to compare against the Youtube Vid and I provided what he asked for!!!!
And Deus Ex Mankind is NOT a next gen title it is rather fubar tbh, in fact Tom's publication has been trying to get answers from the dev on why it is behaving that way; you do not think it strange that 7600k/7700k/6900k all have the same fps within 0.2 fps when they are left to their default core frequency and boost settings?

With regards to memory, I have linked data also showing improvements of memory frequency works equally for both Intel and Ryzen in games as Eurogamer has done this test on both and I posted the data, it is not unique to Ryzen.
Here it is again the request regarding GTA V:



Cheers

Edit:
And unfortunately seems Paul at Tom's used internal benchmark of Deus Ex Mankind Divided, so that really does not help and would raise into questions any result.
Click to expand...
Just using information you provided is all, nothing negative meant for you. How about the Division where the 7700K does beat RyZen but really not by that much showing that the 4 banger can be caught by the V-8 monster at times. Up the ram to 3200 and it would have probably beat the 7700K is the jest. In other words not everything is cut and dry. Long term gaming ability of the 8 core Ryzen to me looks good.
 
noko said:
GTA V is not a future type coded game - yesterday more Intel optimized game. I have no problem with my 3440-1440p maxing out a constant 60fps in this title. No cpu, graphics card at this time can improve my game play because I am VIDEO MONITOR LIMITED. Which is a new concept for many I take. My hardware is driving my monitor at it's fastest speed with room to spare. My eX Fx 9590 could not do that.

Looking at DX 12 newer titles the differences look to become much smaller where for some reason Dues Ex DX 12 for RyZen smokes Intel. Maybe a testing error but I do find interesting. The benchmark below indicates a limitation on both platform but in this case never explored. Neither cpu speed differences made that much overall in the overall benchmark. So what is going on here, why could RyZen process the data faster than any Intel CPU? Is the larger cache of RyZen the reason? Can other titles take advantage of that? I would think so if the case.

View attachment 20548

Now with all of these low resolution gaming bench that is suppose to be testing the cpu - where is the corresponding CPU usage for a given benchmark? Would running a benchmark that the cpu is only using 30% or less capability would even be considered a cpu test? I just don't see it that way. It does hint at other limitations which seems like are never explored but some nonsensical general conclusion such as RyZen sucks at gaming or not best at gaming BS without even trying to dig a little deeper.

Now the ram speed was at DDR 2666, which is the max rated speed for single sided Dim, 2 modules. I am sure the results would be better at DDR 4 3200 but not everyone will be able to achieve that even with the best sticks of ram. Some RyZen IMC's will just not do it.
Click to expand...

Just to clarify
Here is Deus Ex Mankind Divided without the internal benchmark, and performance swaps in favour of Intel.

RYZEN7-1800X-56.jpg


It is closer than it should be because he is bottlenecking the upper results due to using bloody 4xMSAA and even bottlenecking a GTX Titan X Pascal, really should do the CPU test without any type of AA but you do need the highest game option settings (without any AA) as some higher options are done on the CPU.
Anyway shows why one never should use internal benchmark as this flips the Tom's result that did use it.

Cheers
 
CSI_PC said:
Just to clarify
Here is Deus Ex Mankind Divided without the internal benchmark, and performance swaps in favour of Intel.

RYZEN7-1800X-56.jpg


It is closer than it should be because he is bottlenecking the upper results due to using bloody 4xMSAA, really should do the CPU test without any type of AA but you do need the highest game option settings (without any AA) as some higher options are done on the CPU.
Anyway shows why one never should use internal benchmark as this flips the Tom's result that did use it.

Cheers
Click to expand...
Thanks, better more precise testing - So TomsHardware flubbed it up. Who did this? Link please?
 
noko said:
Thanks, better more precise testing - So TomsHardware flubbed it up. Who did this? Link please?
Click to expand...

Ah sorry yeah I should had included link.
http://www.hardwarecanucks.com/foru...-amd-ryzen-7-1800x-performance-review-16.html

I like HardwareCanucks when it comes to games as they do a very good job of data analysing games and never rely upon internal benchmarks; another reason why HOCP is great and also PCGamesHardware and Gamers Nexus as they also refuse to use internal benchmark tools and their data capturing fps performance; AoTS is an example where they may have to use the benchmark to remove as much randomness as possible but the data performance capture is done with external independent tool such as PresentMon.
Tom's is still better when it comes to power measurements/specific tests such as performance-efficiency across frequency and power with GPUs/etc.
However not sure what is going on with Tom's as I find the best two for doing reviews is Chris and Igor, just feels like Paul dropped the ball as something is also different with his TDP results on Intel when he compared them to Ryzen.
Although it was a mistake by HardwareCanucks to use 4xMSAA, that really hammered the GPU and bottlenecked it at the end for Intel but by how much who knows.
Cheers

Edit:
Where sometimes HardwareCanucks does use internal benchmark it is to reduce randomness but they still capture and measure with outside tools, specifically they use PresentMon.
Just realised they may do that for GTA V as it is a random world, but the tool used to capture performance is independent of the game.
PresentMon has been accepted as the best way for now to measure DX12 games.
 
Last edited:
CSI_PC said:
Ah sorry yeah I should had included link.
http://www.hardwarecanucks.com/foru...-amd-ryzen-7-1800x-performance-review-16.html

I like HardwareCanucks when it comes to games as they do a very good job of data analysing games and never rely upon internal benchmarks; another reason why HOCP is great and also PCGamesHardware and Gamers Nexus as they also refuse to use internal benchmark tools in games.
Tom's is still better when it comes to power measurements/specific tests such as performance-efficiency across frequency and power with GPUs/etc.
However not sure what is going on with Tom's as I find the best two for doing reviews is Chris and Igor, just feels like Paul dropped the ball as something is also different with his TDP results on Intel when he compared them to Ryzen.
Although it was a mistake by HardwareCanucks to use 4xMSAA, that really hammered the GPU and bottlenecked it at the end for Intel but by how much who knows.
Cheers

Edit:
Where sometimes HardwareCanucks does use internal benchmark it is to reduce randomness but they still capture and measure with outside tools, specifically they use PresentMon.
Just realised they may do that for GTA V as it is a random world, but the tool used to capture performance is independent of the game.
PresentMon has been accepted as the best way for now to measure DX12 games.
Click to expand...
Yep, that was a good review. I think you should pick one of these up just because events like these don't happen that often anymore. Going from Bulldozer to RyZen is such an incredible jump or more like miracle for AMD. Also OCing is more like the old fashion way where you have to grind away to get somewhere but is rather fun.
 
You must log in or register to reply here.
Back
Top