[PCper]Ashes of the Singularity Gets Ryzen Performance Update

Took around 400 hours to do, not much for a small team really. Since it's an engine update presumably one could see similar gains in UE and Unity etc if the vendors release and update.
 
Well Ryan did try play that up "why no intel improvements". It's simple really, oxide saide they were going to do a Ryzen patch specifically to address performance on Ryzen parts...

Yeah 20% is massive and legit reviews echoed the same. Remember when people said you'll be lucky to get 5% gains in optimization
 
OrangeKhrush said:
Well Ryan did try play that up "why no intel improvements". It's simple really, oxide saide they were going to do a Ryzen patch specifically to address performance on Ryzen parts...

Yeah 20% is massive and legit reviews echoed the same. Remember when people said you'll be lucky to get 5% gains in optimization
Click to expand...

The crux of the argument will be if that Ryzen-only optimization patch shows up for current games and if future releases reflect that optimization. Now looking at PCper's results it's basically less than 5 fps of a 6900k. Also your legitreviews link addresses the concern between 1700, 1700x, and 1800x. A stock clocked 1700 is (off the bat looking) is following closely the performance of a 1800x in AoTS and therefore near 6900k. Very good.
 
This is great and shows what potential is there. Legitreviews had 31% improvement at 1440p. Not sure how many game devs will go back and patch old games, but this may be what we can expect in new titles.
 
LoL. I can't believe how quiet this forum is right now lol. Seems like all the people that said Oh Oxide CEO always says shit and this will never happen. May be 5%. So much for that. This shows that there is such thing as optimization for new titles and intel has dominated the gaming industry so much no really cared about the FX series of processor since they were such a large disappointment or AMD CPUs in general. Now that the IPC is there a few weeks and we a little tweaking we have these large gains. Amazing, I hope other developers will be more involved and AMD is out there grinding along with game developers to make sure they help them optimize games for AMD processor as well.
 
Ryzen is work in progress, the only blame i give was that AMD were way behind the development curve. Intel had the Skylake/Kabylake X parts in final QS last year June,. Napels was unknown by January this year. That is one area AMD need to get better at, stop dragging feet and putting it on board partners or Microsoft or game devs to rush fixes.

Essentially that is my overall wish for AMD.
 
I liked that video, straight to the point and showed actual cpu usage and thread usage, no funny faces, long worthless dialogues etc. Very refreshing.

This I do believe shows that some serious optimizations or potential does exist yet for RyZen cpu's for games. So some very good news.
 
Round 2 or 3 of Zen should be awesome for sure. Get the bugs worked out, get the optimizations out there, and get the manufacturing/architecture improvements to get the clock speed up, and AMD would no longer just be the best value brand. It would be best value and near top, if not at the top performance.
 
I've felt this for a while, but it seems AMD is playing a long game with their architecture choices. It doesn't always pay off immediately but I do believe in the very near future we will see Ryzen deliver on all the promise and more.
 
SighTurtle said:
https://www.pcper.com/reviews/Processors/Ashes-Singularity-Gets-Ryzen-Performance-Update

15-20% improvement in performance is actually more than I expected.
Click to expand...

Sounds about right and great to see them able to do this pretty fast.
But WTF was Ryan doing putting in just a 2400MHz 6900K, for context that also needed 3200MHz along with the slower memory otherwise he should had not put it in and just commented in the article that the changes were Intel neutral.
Some in the comments there and other places are already misreading context of the Intel performance to Ryzen with new patch sigh.
Interesting that going from 2400MHz to 3200MHz does not have a great effect in this game that is heavily CPU orientated.

The one concern though for some of the recent/older releases, seems the Oxide devs cannot do this without help from AMD engineers but it is fair to say Oxide is not as big as AAA studios in terms of resources available.
Fingers crossed the AAA studios can do some prompt patching, may depend how much they can do based upon AMD's level of engagement.
But key point for any decision on existing games and importantly game engines is: the number of hours requried as mentioned in the article, "The result of 400 developer hours of work".
More likely to see changes for existing game engine going forwards and probably only a few existing game, but the positive is it shows it can be done.
Cheers
 
Last edited:
OrangeKhrush said:
http://www.legitreviews.com/amd-ryzen-performance-update-released-ashes-singularity_193137

My guess is AOTS is a crap game will become the mantra
Click to expand...

My primary concern with AoTS/Oxide is that they are heavily influenced by AMD, we saw this with the internal benchmark tool when compared to a more independent 3rd party tool when it was used for comparing GPUs between AMD and Nvidia.
Now in theory that should not influence what we see now with CPUs, but after how the internal tool worked with regards to GPU comparison it is something I am wary of, ideally the CPU benchmark results need to be validated with PresentMon.

Cheers
 
CSI_PC said:
My primary concern with AoTS/Oxide is that they are heavily influenced by AMD, we saw this with the internal benchmark tool when compared to a more independent 3rd party tool when it was used for comparing GPUs between AMD and Nvidia.
Now in theory that should not influence what we see now with CPUs, but after how the internal tool worked with regards to GPU comparison it is something I am wary of, ideally the CPU benchmark results need to be validated with PresentMon.

Cheers
Click to expand...

Why do the CPU results have to be validated, Kyle Bennett just validated them.
 
CSI_PC said:
Sounds about right and great to see them able to do this pretty fast.
But WTF was Ryan doing putting in just a 2400MHz 6900K, for context that also needed 3200MHz along with the slower memory otherwise he should had not put it in and just commented in the article that the changes were Intel neutral.
Some in the comments there and other places are already misreading context of the Intel performance to Ryzen with new patch sigh.
Interesting that going from 2400MHz to 3200MHz does not have a great effect in this game that is heavily CPU orientated.

The one concern though for some of the recent/older releases, seems the Oxide devs cannot do this without help from AMD engineers but it is fair to say Oxide is not as big as AAA studios in terms of resources available.
Fingers crossed the AAA studios can do some prompt patching, may depend how much they can do based upon AMD's level of engagement.
But key point for any decision on existing games and importantly game engines is: the number of hours requried as mentioned in the article, "The result of 400 developer hours of work".
More likely to see changes for existing game engine going forwards and probably only a few existing game, but the positive is it shows it can be done.
Cheers
Click to expand...


Bethesda *cough* Dice
 
OrangeKhrush said:
Why do the CPU results have to be validated, Kyle Bennett just validated them.
Click to expand...

For the very reason I mentioned; all the new patch AoTS tests by sites including here only used the internal tool both for the scenario and the fps....
That is not representative of what actually happens outside of the gaming engine because Oxide measure 'performance' different to traditional independent tools such as FCAT/FRAPs and importantly PresentMon.
Those review sites that used PresentMon on AoTS in the past for GPU performance showed that the performance swapped in favour of Nvidia from AMD, the benchmark tool does not indicate actual user presented frame behaviour and performance.

As I said in theory it should not matter at a CPU level, but considering we have this way of analysing in the tool that favoured AMD historically I would like to see it validated with independent 3rd party tool such as PresentMon not only from the perspective of close working with AMD by Oxide but also their data capturing is slightly skewed compared to actual presented user performance.

Cheers
 
I think where the communication breaks down is the 6900K and 7700K both serve as high water markers, the closer AMD gets to those markers the more they over exceeded expectations.

I also absolutely reject the statement that claims Oxide were somehow coerced to helping AMD, that is flat out vexatious. Oxide is a game developing company that make money selling their games, they make more money if they can sell their title to both AMD and Intel users, marginalizing yourself to one brand loses you income.

Other developers will follow the mandate because they want AMD users buying future titles, doesn't matter if it is magic moose productions or EA games.
 
OrangeKhrush said:
I think where the communication breaks down is the 6900K and 7700K both serve as high water markers, the closer AMD gets to those markers the more they over exceeded expectations.

I also absolutely reject the statement that claims Oxide were somehow coerced to helping AMD, that is flat out vexatious. Oxide is a game developing company that make money selling their games, they make more money if they can sell their title to both AMD and Intel users, marginalizing yourself to one brand loses you income.

Other developers will follow the mandate because they want AMD users buying future titles, doesn't matter if it is magic moose productions or EA games.
Click to expand...

Who said coerced?
This is R&D engine level design changes and PCPer mentioned it involved AMD technical staff, same that any modifications are not done by the usual more broader game devs but by the inhouse specialist team that are core to any R&D and also engine/rendering/physics/etc development.
That is one aspect that any changes needs to consider and how it requires that small R&D high expertise team that exist in the studio as this is at a game engine level focus, hence why I feel not many existing games will be changing on an individual basis but the existing engines for future games probably will be.
Cheers
 
Last edited:
CSI_PC said:
Who said coerced?
This is R&D engine level design changes and PCPer mentioned it involved AMD technical staff, same that any modifications are not done by the usual more broader game devs but by the inhouse specialist team that are core to any R&D and also engine/rendering/physics/etc core development.
That is one aspect that any changes needs to consider, it is time required by the smaller specialist high expertise team in a dev Studio who do R&D engine/rendering/simulation designs.
Cheers
Click to expand...

Future games will incorporate AMD instruction handling into the consideration, the issue is whether current or old titles will share the same attention, it is marginal but irrelevant as in a years time nobody will care about rise of the tomb raider, far cry primal or deus ex, they will give a crap about whatever replaces them, none of those titles share longevity like BF one for instance does and Dice are also working on Ryzen code updates, and were also partnered to AMD during ryzens development.
 
OrangeKhrush said:
Future games will incorporate AMD instruction handling into the consideration, the issue is whether current or old titles will share the same attention, it is marginal but irrelevant as in a years time nobody will care about rise of the tomb raider, far cry primal or deus ex, they will give a crap about whatever replaces them, none of those titles share longevity like BF one for instance does and Dice are also working on Ryzen code updates, and were also partnered to AMD during ryzens development.
Click to expand...

Well they care enough for now with AoTS patch to talk about it with its own thread and that is what my posts are in context of :)
Please note I mention and talk about existing games and existing enginers (this is the key one as they have a longer cycle and used for multiple future games but as I said these engines will more than likely be updated) as they fit with this thread.

Of course new game developments will take Ryzen into consideration (not necessarily all but quite a lot), and as I have mentioned in several places now potentially Windows Game Mode may also benefit Ryzen for existing games (of course needs testing).

Cheers
 
Well, this is awesome. While I'm not gaming as much as I used to, these days, Ashes of the Singularity is one of my favorites. Nice to see it running better on my setup. Between this and DOTA 2... Looks like AMD was telling the truth. Ryzen optimizations help a lot. All I wanted was within 5-10% of Intel on fps... so I could enjoy the productivity benefits of 8 cores. Looks like AMD is delivering, now.

They just need to be very supportive of developers in this, and build some goodwill, so the optimizations actually get done.
 
OrangeKhrush said:
Future games will incorporate AMD instruction handling into the consideration, the issue is whether current or old titles will share the same attention, it is marginal but irrelevant as in a years time nobody will care about rise of the tomb raider, far cry primal or deus ex, they will give a crap about whatever replaces them, none of those titles share longevity like BF one for instance does and Dice are also working on Ryzen code updates, and were also partnered to AMD during ryzens development.
Click to expand...

They will care about these games. Well, maybe, anyway. I still play a lot of old sh*t, when I have the time. But here's the thing... even if games like Rise of the Tomb Raider never get optimized... who cares? Ryzen does fine with these games. The bone of contention with Ryzen's gaming performance wasn't that it couldn't handle modern games at an absolute level (obviously it can)... it was the possibility that Ryzen's relatively low performance compared to Intel was indicative of the future. That an Intel CPU would last much longer in gaming than a Ryzen chip. If optimizations on future titles put Ryzen back into competitive range with the Intel offerings, then Ryzen has proven to be future-proofed in this respect.

The DOTA2 and AotS results suggest that optimization fixes the issue. Thus, the only potential problem now is whether or not AMD can get developers on board to do such optimization for future titles. AMD needs to build the goodwill and market share necessary to make this happen. But this is very promising for them, nonetheless.

I remember when AMD bullsh*tted us with a similar claim back when Crapdozer came out. But optimization got them what, 2% benefit? If that? But with Ryzen we see double digit increases in *both* titles optimized so far. That's very promising.
 
CSI_PC said:
My primary concern with AoTS/Oxide is that they are heavily influenced by AMD, we saw this with the internal benchmark tool when compared to a more independent 3rd party tool when it was used for comparing GPUs between AMD and Nvidia.
Now in theory that should not influence what we see now with CPUs, but after how the internal tool worked with regards to GPU comparison it is something I am wary of, ideally the CPU benchmark results need to be validated with PresentMon.

Cheers
Click to expand...

What is this weird language "influenced" ? AMD has consumer parts that can get us out of this stagnant gaming environment.
DuronBurgerMan said:
The DOTA2 and AotS results suggest that optimization fixes the issue. Thus, the only potential problem now is whether or not AMD can get developers on board to do such optimization for future titles. AMD needs to build the goodwill and market share necessary to make this happen. But this is very promising for them, nonetheless.
Click to expand...

Well the optimizations are "for" the Nitrous engine more then just AotS.
 
DuronBurgerMan said:
The DOTA2 and AotS results suggest that optimization fixes the issue. Thus, the only potential problem now is whether or not AMD can get developers on board to do such optimization for future titles. AMD needs to build the goodwill and market share necessary to make this happen. But this is very promising for them, nonetheless.
Click to expand...

I really don't think this is an issue going forward. Pre-existing games were essentially optimized for Intel architecture, since there was no Ryzen. If we lived in a world were Ryzen was the standard and Intel had been running Core 2 variants until now and just came out with Kaby Lake, chances are there are games which would show some poor Kaby results in games as they would be optimized for Ryzen, and not some new Intel Architecture.

All CPU architectures have some unique quirks, and you have to take these into account to get the best results.

Going forward anyone developing new games will be testing/optimizing on Ryzen as well as Intel, they won't necessarily have the same level of optimization, but the bigger gaps driven by low hanging fruit, should disappear.
 
Pieter3dnow said:
What is this weird language "influenced" ? AMD has consumer parts that can get us out of this stagnant gaming environment.


Well the optimizations are "for" the Nitrous engine more then just AotS.
Click to expand...

Everything Oxide has done aligns with AMD; they were one of the very 1st to go on about Mantle and continued to champion it, same goes for Async Compute suggesting improvement figures we never saw (possible on console but not PCs), the 1st developer to come out and say we can improve our game for Ryzen so they act as a Ryzen tech champion to show it can be done.
Everything including the internal benchmark tool with its frame analysis that if you use it shows better performance for AMD than independent tools that look at performance in the traditional sense as the present frame to user rather than internal to the game engine.....

If that is not influenced I have no frigigng idea what is, because many here would be moaning very quickly that Nvidia paid them off if it was the other way round.
Ask yourself why such an indie company was 1st with AMD's engineers in so many aspects, funnily also with improving AoTS for Ryzen and surely there would be other games-studios that would make sense but then they do not work as closely nor as influenced by AMD as Oxide.

But this is moving away from my actual OP post that was.
Ashes of the SIngularity even in this instance for CPU benchmark testing needs validating by PresentMon, as outlined in my earlier posts.

Cheers
 
CSI_PC said:
Everything Oxide has done aligns with AMD; they were one of the very 1st to go on about Mantle and continued to champion it, same goes for Async Compute suggesting improvement figures we never saw (possible on console but not PCs), the 1st developer to come out and say we can improve our game for Ryzen so they act as a Ryzen tech champion to show it can be done.
Everything including the internal benchmark tool with its frame analysis that if you use it shows better performance for AMD than independent tools that look at performance in the traditional sense as the present frame to user rather than internal to the game engine.....

If that is not influenced I have no frigigng idea what is, because many here would be moaning very quickly that Nvidia paid them off if it was the other way round.
Ask yourself why such an indie company was 1st with AMD's engineers in so many aspects, funnily also with improving AoTS for Ryzen and surely there would be other games-studios that would make sense but then they do not work as closely nor as influenced by AMD as Oxide.

But this is moving away from my actual OP post that was.
Ashes of the SIngularity even in this instance for CPU benchmark testing needs validating by PresentMon, as outlined in my earlier posts.

Cheers
Click to expand...

All fair points. If Dota 2 didn't show similar findings -- i.e. if AotS were the only game showing improvement, I'd be much more suspicious. But we have a situation where Ryzen performs well in a lot of areas, and seems to drop off in gaming... and then AMD claims this is lack of optimization. Both examples we have now of optimization show double-digit improvement. Now, I will grant, this is a small sample size, and one of the examples has an obvious pro-AMD outlook. So this doesn't wrap up the issue with a nice little bow for AMD just yet.

But it *is* very promising, nonetheless.
 
Thanks to lolfail linking this in another thread.
Here is the AoTS patch showed with performance figures internally and importantly measured with independent tool PresentMon.
Using the internal tool completely including for frame performance puts Ryzen at the top, but using the independent PresentMon that monitors at the Present user end rather than game engine-buffer sync showed the 7700K still very much on top.
Hence the internal tool is skewing the data performance of when one really plays, and unfortunately the gains are not as close as one can really expect.
http://www.pcgameshardware.de/Ryzen...ecials/AMD-AotS-Patch-Test-Benchmark-1224503/

It is in German and so needs translating but can still look at the charts without translation, notice how the performance swaps when looking at it with the external performance measurement tool and in favour of Intel.

So the internal tool scenario + frame performance capture (data capture point is the important factor here) skews in favour of AMD, the frame figures to use are the independent ones not provided by the internal tool.
Cheers
 
Last edited:
Man I hope devs jump ALL OVER THIS. And optimize code to recognize and run at optimal performance for multi threaded games. MORE MORE MORE. Soon.. SO SOOON.. Games will actually start using 4 and 8 and more threads well!

Then in 5 years when I upgrade again.. I can get a 32 core 64 thread CPU and install storage that needs 16 pcie channels that is fully predictive!

Ok the storage bit is a joke but a home system with 32 cores and 64-128GB of ram would ROCK.

Throw in the new Vega technology from AMD and you could host multiple instances of your games running on your portable VR devices running in Virtual OS's on your home PC. Ohhh yea baby!

That right there.. is actually possible.. just a matter of getting wireless video up to snuff with wireless networking. You will put on a battery powered thin client for your Virtual gaming environment hosted on the home PC. Then play your games! Or you know.. Pron.
 
CSI_PC said:
Thanks to lolfail linking this in another thread.
Here is the AoTS patch showed with performance figures internally and importantly measured with independent tool PresentMon.
Using the internal tool completely including for frame performance puts Ryzen at the top, but using the independent PresentMon that monitors at the Present user end rather than game engine-buffer sync showed the 7700K still very much on top.
Hence the internal tool is skewing the data performance of when one really plays, and unfortunately the gains are not as close as one can really expect.
http://www.pcgameshardware.de/Ryzen...ecials/AMD-AotS-Patch-Test-Benchmark-1224503/

It is in German and so needs translating but can still look at the charts without translation, notice how the performance swaps when looking at it with the external performance measurement tool and in favour of Intel.

So the internal tool scenario + frame performance capture (data capture point is the important factor here) skews in favour of AMD against Nvidia GPUs (need to see it in specific GPU reviews not this article as 'GPU Benchmark' option in AoTS is relevant still to CPUs), and also against Intel for CPUs.
Cheers
Click to expand...

https://translate.google.com/transl...ch-Test-Benchmark-1224503/&edit-text=&act=url
 
Well how about this? I will run the test *myself* on my machine, and tell you the results, both before and after. Supposedly this version is in the wild on Steam already. I already ran the benchmarks on the old version last night... so I will run them again and give both PresentMon and Internal figures, and then also give you a subjective analysis (like is it jerky or not).

I play this sh*t all the time, so I'll know right away if there is funny business.
 
DuronBurgerMan said:
Well how about this? I will run the test *myself* on my machine, and tell you the results, both before and after. Supposedly this version is in the wild on Steam already. I already ran the benchmarks on the old version last night... so I will run them again and give both PresentMon and Internal figures, and then also give you a subjective analysis (like is it jerky or not).

I play this sh*t all the time, so I'll know right away if there is funny business.
Click to expand...
The update rolled out about 9pm CDT, so unless you have an offline version from before then, testing the old version will be tough. Just an FYI for those looking to do their own comparisons.
 
CSI_PC said:
Everything Oxide has done aligns with AMD; they were one of the very 1st to go on about Mantle and continued to champion it, same goes for Async Compute suggesting improvement figures we never saw (possible on console but not PCs), the 1st developer to come out and say we can improve our game for Ryzen so they act as a Ryzen tech champion to show it can be done.
Everything including the internal benchmark tool with its frame analysis that if you use it shows better performance for AMD than independent tools that look at performance in the traditional sense as the present frame to user rather than internal to the game engine.....

If that is not influenced I have no frigigng idea what is, because many here would be moaning very quickly that Nvidia paid them off if it was the other way round.
Ask yourself why such an indie company was 1st with AMD's engineers in so many aspects, funnily also with improving AoTS for Ryzen and surely there would be other games-studios that would make sense but then they do not work as closely nor as influenced by AMD as Oxide.

But this is moving away from my actual OP post that was.
Ashes of the SIngularity even in this instance for CPU benchmark testing needs validating by PresentMon, as outlined in my earlier posts.

Cheers
Click to expand...
AMD Stardock.png
 
DuronBurgerMan said:
It's funny... I didn't realize this game was so unpopular. I love it.
Click to expand...
Last year in February it wasn't so bad with a peak of 700 players and 300 average. At least when I climbed the ladder I wasn't going up against bots all the time, or waiting 10 minutes for a match (just to go against a bot). Problem was people kept comparing it to Starcraft. In a few years the game will be better I hope; at least if I go by the fans of Sins of the Solar Empire.
 
You must log in or register to reply here.
Back
Top