AdoredTV - Low resolution benchmarks are worthless

Bit of a storm in a teacup.

I think it's fair to say that AMD are marketing towards gamers and why wouldn't they, that's a significant part of the market. I think It's also fair to say that these initial releases are not honestly conceived as gaming chips, gaming is just likely use case.

Ryzen is looking pretty fantastic to me - even for gaming. I'm not currently in the market for a CPU but if I was I wouldn't have any hesitation in lining up any of the products released so far. People always shoot for the moon where performance is concerned but this isn't always the most useful measurement. Not all use cases and set ups are equal. Eg I'm still on a 60Hz monitor and will be for at least a year. So the fact Ryzen can only pull ~120fps in GTAV is irrelevant. The fact it would allow for multiple background VMs, great productivity and (presumably) streaming on the other hand means a lot. Same product different perspective.

Let's wait for the 4 and 6 core parts and see what happens.

3 Likes

There we go...

1 Like

I'm excited for the 6 cores as well. Do we have any information on release date for the 4 and 6 core skus?

the secret, is that adored is raja. (joking, but maybe - maybe not)
now that we sew some doubts lets take a look at this thread.

(didn't watch the vid)

yes low res game tests are worthless, as they don't reflect performance, but point of bottleneck with certain workloads. The most important cases, are ones with current types of work you expect it to do, and you are not expected to run games at those rez anyway, you won't be getting 300fps normally... while its great that we can see bottleneck somewhere it doesn't tell us anything anyway... it could be memory controller, windows scheduler, cache access bugs, you name it - it could be the issue)

Best way to benchmark cpu, is to give it expected type of work. 1080p, 1440p, 4k
(people started to move away from 1080p, but at least 2 more gens of gpu's before 1440p or 4k is like 1080p today.) If they are getting edge over intel at higher resolutions that also means something (not sure if they did) so those results cannot be discarded; and testing for longevity of cpu is best tested by its flops. raw power. If it has a lot of it, then expect it to survive. If not, well its going to be a crisp in 2yr or so...
(don't expect you're going to get higher clocks from your cpu's, likely its the best they can do... but motherboard could be ones to blame too - and i assume 60% of problems with performance could be given to poor early mobo implementations and bugs.)

you will not see better results from 4-6 core parts... forget it. when you are cutting you are cutting.
Only thing they could have less is latency with memory, (ex. i5-2500k had the lowest latency of ~20-30ns while i7-2600k had around 65ns)

1 Like

Everybody has their own opinions on the presented benchmarks.
To me every vorm of test and benchmarks can be interesting, because they all tell a story.

1 Like

AMD during the Polaris reveal about how IMPORTANT 1080p performance is:

AMD during the Ryzen review cycle when their chips frankly don't perform as well as Intel chips at 1080p currently:

"1080p isn't a good test scenario" (I don't have the exact wording to hand but you get it).

I will note that it is still accurate that >95% of steam users have 1080p monitors or lower, making reviews that occur right now still completely valid to test at that resolution. Currently 43.23% of steam users are on 1080p, and 1080p is also the fastest growing resolution in the survey having gained 4.13% from the time the last survey was conducted. The total percentile of single monitor solutions above 1080p is 4.34%, with a net 0.43% growth from last quarter in monitor resolutions greater than 1080p.

I think reviews and recommendations probably make the most sense if they are conducted with the information at hand and not relying on the caveat that "It'll get better". Yes, it will get better, but its a reviewers job to review the product that sits in front of them, not the hypothetical product it will be in 1 or 2 years. They should be mentioning that, "Hey Ryzen isn't as good as the intel competition as it sits right now, but it'll get better." But, at the end of the day, they're recommending a product that is good now, not later.

I will also note that I think it really hurt Adored's argument to use Joker's benchmarks. Regardless of if you think Joker tampered with the results (I don't think he did but that's not the topic at hand), you must admit that his results are still a massive outlier from the rest of the other reviewers in the space. Whether that have been caused by other programs running on that machine, or windows update, or whatever, doesn't matter. All that matters is that something was wrong with his numbers due to some random factor, and that using those numbers for a basis of an argument is not wise if you want to be taken seriously by me or others that distrust Joker's results.

Man Raja can make a sick Scottish accent.

nah, he doesn't have Scottish accent.

this is Scottish accent.

1 Like

Give him a break...It is a big leap from his native one...

i'm not complaining about his hindu accent.

But you are not appreciating the effort to go from one to the other..

hmm, i couldn't care less really.

pretty sure this is just AMD's #1 fan doing rationalization and damage control because the launch performance was more realistic than he had hoped

1 Like

You really believe those 95% 1080p gamers will buy 8core Ryzen? You don't think, that the 8 cores are kinda made for an entirely different market segment? Maybe the quad cores for 150-160$ would be more representative of what 95% 1080p gamers with gpus under 200$ may actually buy for gaming and not the 499$ 8core?

People buy 7700k, 6700k, 6800k, and 6850k's to play 1080p all the time. Look at the highest rated builds for this 1080p monitor on pcpartspicker: https://pcpartpicker.com/product/Nbbp99/acer-monitor-umvh6aa003

First build, 6850k.
Second build, old.
Third build, 6700k.

That's the second highest reviewed 1080p monitor on the site. The first is 144hz, this one is 60. Keep in mind the 1700 is a direct price competitor with the 7700k, and the statistics don't lie that 1080p is still the dominate resolution by a landslide.

Just as much as the RX 480 was a "budget" oriented gpu in the sceme of things, the 1700 isn't that much farther off in the landscape of cpus. The RX 480 was a $250~ gpu in a market with a $1200 consumer oriented gpu, and the 1700 is a $330 cpu in a market with a $1500-1600 consumer oriented cpu.

Its also not that uncommon for people to not upgrade monitors with just a cpu + mobo upgrade. There are a lot of people that were holding out on AM3+ to move to Ryzen. There were about 300 of them on this forum alone that would all mention it every time Ryzen news came out and they got excited. Not all of those people did a complete system upgrade along with gpu and monitors. The percentile of steam users using 1080p monitors makes it likely that this usecase isn't all to uncommon percentile wise within the ryzen 7 user base.

Either you don't get my point or I don't get yours...
I still think for the price it's more than amazing processor... Outperforming twice as expensive chips in some applications and not in others, but it's half the price...
It's a brand spanking new architecture, so the only way is up, and I still don't believe people will buy 8 core 16 thread cpu for 1080p gaming.
Why? Because I am waiting for the 6 cores for editing purpose, otherwise I would get 4c8t and call it a day.
1080p is not irrelevant resolution in general. It's irrelevant in the specific case.
And to overblown the situation and ignore all other benchmarks a d say it's a failure because of 1080p is laughable for the tech media...

I have a horrible feeling that those 8 core parts were solely to tempt apple away from intels overpriced chips... doubt gaming performance will increase in leaps and bounds in the next year as there is still too much fuckery going on with bullshit dx12 releases made by monkeys.

...If I was apple I would be considering them to replace the overpriced intel parts :D

Most users, are still on 1080p. That means, that people that are doing a pure cpu+mobo+ram swap are most likely to be swapping into systems using 1080p monitors, because that is the most common resolution percentile wise. Perfect examples of this are the only two Ryzen builds on pcpartspicker with a monitor listed. Both did a straight cpu+mobo+ram swap. One is using 1920x1080 at 144hz, and the other is 1920x1200 at 60hz. This is really a prime demographic for the 1700 to capitalize on.

https://pcpartpicker.com/b/fQV6Mp

http://pcpartpicker.com/b/nzMZxr

There are most definitely more people out there like this running 1080p that are upgrading to Ryzen 7 chips. The percentile chances make it so that there will undoubtedly be users that meet this description, just as the two above builds have. Quite frankly its a bit ridiculous to me to be disregarding 1080p users as non-ryzen 7 oriented consumers. There are plenty of people using 1080p and enthusiast class hardware. I use a 980 Ti and dual 1080p monitors, because I have a 144hz and I like this resolution, size, refresh rate, and price combination. The 1700 is quite frankly aimed at people like myself, that were eyeing a 7700k price range cpu+mobo+ram upgrade.

Furthermore, I think it's also ridiculous to be saying that AMD didn't market Ryzen towards gamers when they've clearly said "gamers" in the Rzyen conferences over and over again and showing gaming benchmarks on their Ryzen rigs, and, lastly, comparing Ryzen to Intel's four core i7s. It's unequivocally marketed towards gamers.

3 Likes

"350fps and 550fps without computers" You cant mean there is no perceivable difference in any computer game between those frame-rates since that's clearly not the case.

Fast moving objects look better the higher you go and the difference is easily seen.

Did you mean something else?