Ashes of the Singularity gets AMD Ryzen performance update

[chinf]

So according to your figures the 1700X is a more expensive CPU.

The cost of the motherboard is a valid and important consideration, unless you somehow get all your motherboards for free?

[elites2012]

im glad this review was made. surprisingly how people think games "just" run better on intel, not knowing that intel and AMD have to send out dev kits for games to utilize their chips perks properly. a lot of gaming companies chose not to use the fx series gaming kits. which made the fx line look bad.

[Xlucine]

So according to your figures the 1700X is a more expensive CPU.

You did not specify the CPU cost in your original post, only that the 7700K "works out cheaper than a 1700X". How would the cost of choosing between a 7700K and a 1700X not work out to include the cost of a suitable motherboard?

[mtyson]

AMD has uploaded a blog post on the optimisations.
And there's a video chat featuring Stardock CEO Brad Wardell discussing game optimizations for the AMD Ryzen CPU.

[CAT-THE-FIFTH]

It seems PCGH used their own test sequence instead of the built-in benchmark:

http://www.pcgameshardware.de/Ryzen-...hmark-1224503/

It seems the improvement in their test sequence is less than the benchmark.

[anselhelm]

I think there is some merit to what he said - yes, the 7700K is the fastest gaming CPU.. yet it somehow works out cheaper than a 1700X or 1800X. That AMD is a smaller team, and on a smaller R&D budget, should mean a cheaper CPU, not a more expensive one.

As rightly pointed out, gaming is far from the only story with these chips, but on that particular point things have to be put in perspective.

I was contesting the idea that being "on par" with a 7700K is somehow something bad, which is what was implied.

Ryzen is amazing and managing to get "on par" with a 7700K in any game is simply outstanding.

[anselhelm]

You mean the 7700K that's the fastest gaming CPU ever made?!

Seriously, you need to look at the facts again and draw a different conclusion.

maybe you need to....we are talking about AotS here, the only game we know of with the ryzen optimisations in....

No, I stand by my comments in light of the evidence beyond a benchmark for a single game.

To date, the 7700K is the best overall chip for gaming. This isn't my viewpoint but the viewpoint of many large, professional processor reviewers.

There are specific games where more cores are used, allowing Intel HEDT chips to come out on top, but they're currently few and far between.

What I am contesting here is that you seem to be implying that being "on par" with a 7700K is bad: it's not. AMD has done an amazing job to build a brand new CPU which can actually compete again.

It seems like you were expecting Ryzen to both outperform Intel's highly-clocked, fewer-cored processors as well as its HEDT lower-clocked, higher-cored processors, which is why I find your expectations unrealistic.

If I have misunderstood your points, please let me know and I will retract this statement.

I respect that we disagree here and bear you no personal ill will.

[shaithis]

A new chip is outperformed by Intels chip by 20% when they have virtually identical clocks.

This is how BD started and look how that ended....yes, they can compete FOR NOW. Intel will just turn the heat up again though. I'd still pay the extra right now for stability and not being at the mercy of devs optimisations though. No point having something be cheaper when it has issues.

I was getting really excited about this release and have just kept getting more and more deflated since release. Enjoy your video transcoding if that's your bag, I'm a gamer.

[scaryjim]

... This is how BD started ....

Not to counter your whole point, which has some merit, but this is NOT how BD started. BD was much, much slower in lightly threaded tests (its PiFast performance was truly embarrassing) despite having a clock speed advantage over the extant Intel chips.

Ryzen is much closer in lightly threaded tests, at a lower clockspeed than the i7 7700k. And unlike Bulldozer, which got very little dev support, Ryzen is already getting performance enhancing patches that actually work. It's getting beaten by 20% in lightly threaded tests - by a chip clocked 10% faster that's had 4 - 5 generations of iterative improvement.

So, will Intel turn up the heat? Well, they haven't bothered for the last 5 generations because they didn't have anything to compete against. They might be able to, but their products for the next couple of generations will already be going through final certification and taping out - they literally can't just come up with a new architecture in the next few months and push it out the door. There's no indication that Intel have been developing a new underlying architecture for the last few years, either. So they're probably stuck with iterative improvements to the existing one, and their return on those improvements has been tailing off for the last couple of gens.

Meanwhile AMD are making iterative improvements to a very new architecture, which should give them more scope for performance tweaks. You only have to look at Steam Roller and Excavator to see what a well tuned iteration can provide without making any major revisions to a core. Over the next couple of years I'd expect AMD's processors to advance faster than Intel's, and I'd also expect to see software tuning add further improvements to performance under Zen.

So no, this is not how Bulldozer started. It had a much higher performance deficit to make up, was already clocked higher, and was competing against a relatively new architecture revision with more scope for iterative improvement. Ryzen is much closer in performance, can attribute some of that difference to a lower clock speed, and is competing with a much older architecture that appears to be approaching the limits of its optimisation. There are all sorts of reasons to be positive and excited about Ryzen...

[shaithis]

So, will Intel turn up the heat? Well, they haven't bothered for the last 5 generations because they didn't have anything to compete against.

I'm starting to think this is a misconception. Funny how according to many (and I was starting to believe it), Intel have not innovated and have actively held back....yet here we are with AMD releasing Jim Kellers latest child and it's still behind Intel. Maybe Ryzen2 will show more than 5% improvements, who knows...all we can do is look at what is here now and there are so many parallels to BD that I am feeling a bit pessimistic about the whole affair.

[DanceswithUnix]

... all we can do is look at what is here now and there are so many parallels to BD that I am feeling a bit pessimistic about the whole affair.

BD was bug ridden and missed performance target, PD fixed that and was usable. If there are similar bugs in Ryzen, then Ryzen 2 could be a corker. I suspect there probably aren't though.

[kompukare]

So it seems PCGH's earlier tests were not valid and they managed to find out what's going on.
In fact, it looks like their 'do our own benchmark' will never work for AotS since the game automatically reduces the settings for any CPU with less than 6 physical cores making it appear that quad core CPUs are much faster than they really are.
The built-in benchmark doesn't do this which means benching with a savegame like PCGH were doing doesn't work when the core count differs so much. So they added a simulated Ryzen 5 1500X (4C/8T).
http://www.pcgameshardware.de/Ryzen-...hmark-1224503/

[CAT-THE-FIFTH]

So it seems PCGH's earlier tests were not valid and they managed to find out what's going on.
In fact, it looks like their 'do our own benchmark' will never work for AotS since the game automatically reduces the settings for any CPU with less than 6 physical cores making it appear that quad core CPUs are much faster than they really are.
The built-in benchmark doesn't do this which means benching with a savegame like PCGH were doing doesn't work when the core count differs so much. So they added a simulated Ryzen 5 1500X (4C/8T).
http://www.pcgameshardware.de/Ryzen-...hmark-1224503/

So basically using the game as a benchmark is pointless since you are not comparing apples to apples.

[kompukare]

So basically using the game as a benchmark is pointless since you are not comparing apples to apples.

Well, I guess that's why it has a built-in benchmark. But then again PCGH might one of the few sites to even have attempted to run their savegame benchmark.

Would be nicer if they showed this in the in-game settings:
"Warning! You have less than 6 physical cores, and AotS will subsequently lower your settings (even though they still say High-Preset). If you wish to force a benchmark run, click this checkbox. This setting will default to back to off the next time you run AotS on a <6 core CPU. Thank you very much, the AotS team."

Wonder if there are any other games which behave like this?

[msroadkill612]

And thats on Nvidia. Imagine the synergies when the equally~ revolutionary vegas is paired with zen?

[kalniel]

And thats on Nvidia. Imagine the synergies when the equally~ revolutionary vegas is paired with zen?

There aren't any synergies.