[Techspot] AMD Ryzen 5 1600 vs. Intel Core i7-7800X: 30 Game Battle! [Links Fixed - Updated]

[Zucker2k]

They test at default settings (so, turboed) and at fixed max OC. Remember, 7820x turbos to 4.3Ghz and Turbo Max is 4.5Ghz. Do you see the fps delta now?

Are you confusing the 7800x and the 7820x?

 

[MarkPost]

you confused me

Techspot takes the cake on that one. They basically clock the Ryzens as high as they'll go and try to pretend they're doing a clock for clock comparison. For what? Either you run the chips at stock, or you run them overclocked, to whatever clocks they can achieve. If you want to test a Ryzen chip at 4Ghz, you go for the 1800x which'll give you 4Ghz (XFR), but the price will be higher and it won't look too good so what do you do? You take the lower priced Ryzen, overclock it to 1800x+XFR levels (the highest you could clock Ryzen without exotic cooling) and test the chips in as few cpu-demanding games as possible. FYI, the 7820x turbos to 4.3Ghz and Turbo Max is 4.5Ghz. http://www.anandtech.com/show/11550...ew-core-i9-7900x-i7-7820x-and-i7-7800x-tested so 4Ghz is actually an underclock, for stock conditions! The narrative is changing. Soon, a clearer picture will emerge.

 

[CHADBOGA]

Thats your opinion. Do you have proof of what you say ? If not then don't harass those who do not believe as you,

Who is he harassing?

 

[krumme]

The cpu total throughput is not the limit in any of the games tested eg on the techspot portfolio. Sans outliers like bf1 that issues 10 threads.
What we have is how api run un different engines. The top 3 and buttom 3 of crappy dx12 implementation could as well be opposite for ryzen vs sklx.
A patch for the game and its upside down.
A properly coded game like bf1 sits right in the middle with about same performance. As expected.

We are mosly looking at how code is interfering with different cache and memory implementations and speed.

 

[Osjur]

This is so goddamn funny when people are saying that these reviews are flawed and Skylake X should be this and that much faster than Ryzen and then these guru's link to reviews where Ryzen is running with memory from 2133 to 2666mhz with NO INFO about what the timings were. And of course Skylake X is usually running with about 1ghz memory speed advantage.

I have newsflash: Ryzen @ 3ghz with 3466mhz C14 and tightened sub-timings is faster than Ryzen @ 4Ghz and 3200 C14 "standard" preset.

Just for giggles: 3dmark API test: 3466mhz C14 + subs vs 2133mhz:
http://www.3dmark.com/compare/aot/223462/aot/223459

Yep, its the same processor & setup

 

[Anarchist Mae]

Techspot takes the cake on that one. They basically clock the Ryzens as high as they'll go and try to pretend they're doing a clock for clock comparison. For what? Either you run the chips at stock, or you run them overclocked, to whatever clocks they can achieve. If you want to test a Ryzen chip at 4Ghz, you go for the 1800x which'll give you 4Ghz (XFR), but the price will be higher and it won't look too good so what do you do? You take the lower priced Ryzen, overclock it to 1800x+XFR levels (the highest you could clock Ryzen without exotic cooling) and test the chips in as few cpu-demanding games as possible. ...

It was a 30 game test, just because the end results don't suit your biases does not make it invalid.

Now talking about the scorched pins, that is a valid point. We do not know if it affected the benchmark scores, a couple of the tests should be repeated with different hardware to verify them. I'm not holding out hope that Techspot will do that.

Edit: To be clear, I think it's actually unlikely that the scorching had any affect on performance given how the 7820X also performed below expectations. It's more likely that a serious hardware failure was narrowly avoided than the CPU was running fine for a long period of time and not crashing.

 

[Zucker2k]

It was a 30 game test, just because the end results don't suit your biases does not make it invalid.

Now talking about the scorched pins, that is a valid point. We do not know if it affected the benchmark scores, a couple of the tests should be repeated with different hardware to verify them. I'm not holding out hope that Techspot will do that.

Edit: To be clear, I think it's actually unlikely that the scorching had any affect on performance given how the 7820X also performed below expectations. It's more likely that a serious hardware failure was narrowly avoided than the CPU was running fine for a long period of time and not crashing.

How did it perform poorly? Half of those tests are gpu limited anyway, yet it managed to fare better than the Ryzen at same clocks of 4Ghz. I'm scratching my head as to why this guy would even bother to overclock the 7820x to 4.5Ghz for 1fps gain? Are you serious? I'm yet to see a more incompetent reviewer on the planet.

 

[Anarchist Mae]

How did it perform poorly?

Only in $/performance did it perform poorly.

Half of those tests are gpu limited anyway, yet it managed to fare better than the Ryzen at same clocks of 4Ghz. I'm scratching my head as to why this guy would even bother to overclock the 7820x to 4.5Ghz for 1fps gain?

Because it demonstrates that for these 30 games at settings people will play at that there's no point buying a 7800X so long as you're comfortable overclocking.

Are you serious? I'm yet to see a more incompetent reviewer on the planet.

You should not call a reviewer incompetent because their test does not fit your biases.

 

[Anarchist Mae]

I'll attempt to explain why the methodology of this test is actually useful, even if the results are not particularly exciting.

Say this test was of 30 CPU bound games, what would that test tell you? It would tell you that the 7800X is faster than the 1600 in CPU bound games. However you'd be missing out on a large part of the picture because your test ignored games that are not CPU bound.

Which is why we have this test, it contains some CPU bound games and some GPU bound games, and some games that are just damn slow for no good reason. This makes it a better test than your CPU only bound test because it covers a wider range of possibilities.

And guess what! It even shows that the 7800X is faster than the 1600 in CPU bound titles, for example you can check the benchmark summary and see that the 7800X was 16 percent faster than the 1600 while playing Total War: Warhammer.

But it also shows that overall it barely matters which one you buy, because one CPU is faster in some titles and the other is faster in others.

 

[Crumpet]

I'll attempt to explain why the methodology of this test is actually useful, even if the results are not particularly exciting.

Say this test was of 30 CPU bound games, what would that test tell you? It would tell you that the 7800X is faster than the 1600 in CPU bound games. However you'd be missing out on a large part of the picture because your test ignored games that are not CPU bound.

Which is why we have this test, it contains some CPU bound games and some GPU bound games, and some games that are just damn slow for no good reason. This makes it a better test than your CPU only bound test because it covers a wider range of possibilities.

And guess what! It even shows that the 7800X is faster than the 1600 in CPU bound titles, for example you can check the benchmark summary and see that the 7800X was 16 percent faster than the 1600 while playing Total War: Warhammer.

But it also shows that overall it barely matters which one you buy, because one CPU is faster in some titles and the other is faster in others.

Yup, buy whichever one suits your gaming needs, and fits your personal budget.

But this isn't a clear case of "Yeah but X is better/worse", and that's OUR point. Ryzen is a very competitive product, and that is evident in fair reviews.

 

[Gikaseixas]

The fact that we're having discussions on who's faster between Intels latest i7 vs AMD's Ryzen 5 (yes five, not seven) CPU is pretty telling. I'm actually looking forward to i9 vs 1900X Threadripper

 

[Zucker2k]

Only in $/performance did it perform poorly.

Because it demonstrates that for these 30 games at settings people will play at that there's no point buying a 7800X so long as you're comfortable overclocking.

You should not call a reviewer incompetent because their test does not fit your biases.

I'd like to see what you have to say when we see finally see the Threadripper gaming benchmarks. Not when there are more than 30 games in the world, some more demanding, some favouring high frequency single core. With Ryzen pegged at 4Ghz, how would it fare in those scenarios? You talk about my bias, but I'm not the one selling an outlier review as gospel, as thread title boldly proclaims. The 4Ghz 7820x bested the 4Ghz 1700 chip in those limited tests overall (no video encoding even, and the plethora of tests the same site showed in the SKL-X review), but TS boldly puts 4.5ghz in the thread title.

Edit: What's even more shady, only 20% of Ryzen R7 1700s can even make it to 4Ghz. @ 1.440v!! So you'll need something close to a golden sample to break 4Ghz. Haha

https://www.reddit.com/r/Amd/comments/5xybp7/silicon_lottery_ryzen_overclock_statistics/

 

[Anarchist Mae]

I'd like to see what you have to say when we see finally see the Threadripper gaming benchmarks. Not when there are more than 30 games in the world, some more demanding, some favouring high frequency single core. With Ryzen pegged at 4Ghz, how would it fare in those scenarios?

Here, I'll make a prediction, Threadripper won't make as much sense for gaming because it's more expensive and offers a lot more cores than even currently good multi-threaded games use.

You talk about my bias, but I'm not the one selling an outlier review as gospel, as thread title boldly proclaims.

This test is not 'gospel', it's just a test that demonstrates the overall gaming performance of two CPUs, yes there are games that do better on either CPU and those are clearly shown in the review.

The 4Ghz 7820x bested the 4Ghz 1700 chip in those limited tests overall (no video encoding even, and the plethora of tests the same site showed in the SKL-X review), but TS boldly puts 4.5ghz in the thread title.

If you're fine spending a lot more money to get on average only a few FPS more than that's fine, but $/performance wise it's clear which CPU is better overall.

 

[Crumpet]

I'd like to see what you have to say when we see finally see the Threadripper gaming benchmarks. Not when there are more than 30 games in the world, some more demanding, some favouring high frequency single core. With Ryzen pegged at 4Ghz, how would it fare in those scenarios? You talk about my bias, but I'm not the one selling an outlier review as gospel, as thread title boldly proclaims. The 4Ghz 7820x bested the 4Ghz 1700 chip in those limited tests overall (no video encoding even, and the plethora of tests the same site showed in the SKL-X review), but TS boldly puts 4.5ghz in the thread title.

Edit: What's even more shady, only 20% of Ryzen R7 1700s can even make it to 4Ghz. So you'll need something close to a golden sample to break 4Ghz. Haha

https://www.reddit.com/r/Amd/comments/5xybp7/silicon_lottery_ryzen_overclock_statistics/

I've only tested about 30 of each SKU at this point, and they all had S/Ns very close to each other so the numbers might change as we continue to test over the next couple weeks.

Even so that still says 70% at 3.9Ghz, which is only 100Mhz less.

 

[Zucker2k]

Here, I'll make a prediction, Threadripper won't make as much sense for gaming because it's more expensive and offers a lot more cores than even currently good multi-threaded games use.



This test is not 'gospel', it's just a test that demonstrates the overall gaming performance of two CPUs, yes there are games that do better on either CPU and those are clearly shown in the review.



If you're fine spending a lot more money to get on average only a few FPS more than that's fine, but $/performance wise it's clear which CPU is better overall.

Don't forget, a cpu does more than game tho, like video encoding (oops, omitted, I wonder why?), and a plethora of other activities where the more expensive cpu would definitely stretch the gap. It'll also last longer because it'll leverage performance capabilities across all applications, be it games or otherwise.

 

[Zucker2k]

Even so that still says 70% at 3.9Ghz, which is only 100Mhz less.

Which is even higher than I've observed. The majority of 1700 chip overclocks I've observed on forums are around 3.8Ghz and lower. 4Ghz @ 1.440v and Asus Real bench stable is quite the achievement.

Edit: You do have Ryzen, right? Try the Y-Cruncher avx stress test to see where your true clocks should be.

 

[Crumpet]

Which is even higher than I've observed. The majority of 1700 chip overclocks I've observed on forums are around 3.8Ghz and lower. 4Ghz @ 1.440v and Asus Real bench stable is quite the achievement.

Edit: You do have Ryzen, right? Try the Y-Cruncher avx stress test to see where your true clocks should be.

The only people i've seen not hit over 3.8Ghz on Ryzen, ANY Ryzen, are the ones using the stock cooler.

If my system is Prime 95 stable running 4Ghz @1.3v then why should I run Ycruncher?

 

[Anarchist Mae]

Don't forget, a cpu does more than game tho, like video encoding (oops, omitted, I wonder why?), and a plethora of other activities where the more expensive cpu would definitely stretch the gap. It'll also last longer because it'll leverage performance capabilities across all applications, be it games or otherwise.

Seriously? Are you just being wilfully dense?

The guy literally tested from scratch two systems, in 5 configurations, with 30 games and at least 3 tests per game per configuration. That's 450 individual benchmark runs. The reason there are no productivity benchmarks in this review is because it's not a productivity review.

 

[TheELF]

exactly..... well I mean, if you plan to play the very same games in 18/19 that you suggest to test now... otherwise the logic is flawed

We are still playing a lot of new games that are based on 11 year old PS3 technology....
Thankfully Nvidia's threaded optimisation extracts a bit more performance in some games like dishonest 2 but in dx12 games for example like DeusEx:MD it's pure 2006...
Three main threads and that's it the rest is not even worth mentioning.
So yes I do believe that for the next couple of years it's pretty safe to say that we will be playing the same games CPU wise we do today.

 

[Zucker2k]

Seriously? Are you just being wilfully dense?

The guy literally tested from scratch two systems, in 5 configurations, with 30 games and at least 3 tests per game per configuration. That's 450 individual benchmark runs. The reason there are no productivity benchmarks in this review is because it's not a productivity review.

Easy there! Made a mistake. No need for the insult. Thought I was in the 7820x vs 1700 thread.

 

[Anarchist Mae]

Easy there! Made a mistake. No need for the insult. Thought I was in the 7820x vs 1700 thread.

No worries, no wonder it seemed crazy

 

[Zucker2k]

The only people i've seen not hit over 3.8Ghz on Ryzen, ANY Ryzen, are the ones using the stock cooler.

If my system is Prime 95 stable running 4Ghz @1.3v then why should I run Ycruncher?

Y-Cruncher because it'll take your chip to new stress-testing level; more than P95 has been able to do with Ryzen so far? By the way, did you see those power consumption numbers for the 1700 at 4Ghz? In Y-Cruncher, a Ryzen chip was pulling 250W at 4Ghz.

 

[Crumpet]

And how much does a 7800X or 7820X pull at 4.5Ghz+?

 

[Anarchist Mae]

Y-Cruncher because it'll take your chip to new stress-testing level; more than P95 has been able to do with Ryzen so far? By the way, did you see those power consumption numbers for the 1700 at 4Ghz? In Y-Cruncher, a Ryzen chip was pulling 250W at 4Ghz.

Are there any benchmarks that compare these two with Y-Cruncher? Would be surprised if a 1600 or 1700 pulls more than 7800X or 7820X.

 

[Crumpet]

Are there any benchmarks that compare these two with Y-Cruncher? Would be surprised if a 1600 or 1700 pulls more than 7800X or 7820X.

Well considering TH have 7820X pulling 180.5w at stock on Prime95.....