AMD Zen 2 (Ryzen 3000) - *** NO COMPETITOR HINTING ***

[nashathedog]

It's because of a 12 core engineering sample that ran at over 5ghz.

Or at least rumours of the existence of it, from AdoredTV.

It's amazing how often nothing concrete balloons into a monster causing people's imagination to run away with them leading to feelings of disappointment when the truth's known, I wouldn't be surprised if the competition (in this case Intel) are behind a lot of the fake news we hear about, that said Raja Koduri liked to throw fuel on the fire of Radeon products causing wild claims of super fast performance so it's likely they're all at it, even for their own products.

 

[chrcoluk]

I dont know how credible those bench results are for the game, but intel has lower ram latency (ram latency very important in games based on my own testing), and certian instructions still have better IPC on intel.

As I said before also cinebench does not equal game performance, I predicted the zen 2 chips would on average be behind intel in games but with a much closed up gap. So if thats the case and I buy zen2 I wont be disappointed it will just be what I expected.

 

[Vish Petrol]

Intel sponsored the 720p testing.

LOL

 

[Drollic]

That leaked review isn't all that promising for gamers. There's no disputing the value as an all rounder though.

 

[robfosters]

t's niggling me now though that the 9900 and a good Z390 board is so much cheaper than 3900 and X570.

 

[N19h7m4r3]

Even without playing at this low resolution, it's important to benchmark in CPU bottleneck scenario to estimate the future proof potential in few years when games will be more demanding on CPU even at higher resolution.

Everyone said the same when Ryzen launched, and kept recommending the 7600k over the 1600. Now the latter is overall better in games also.

I find 480 and 720 testing pointless, it’s about as synthetic as it gets. As it doesn’t account for new game engines, more multi threading support, and developers developing on and with Ryzen to optimize for the architecture.

It’s clear that older, and some specific games still struggle with Ryzen, such as ancient Arma, and FarCry which doesn’t seem to run well on the AMD architecture at all, even if the IPC and clocks are identical. Those are outliers; which is why we need more than just 5-6 games tested at a resolution that 16% of all steam users even run ( most likely low end systems and laptops ).

 

[humbug]

Small game selection, but looking at minimums, they look nice.

Can’t say I give a rats about 480 and 720p testing. Give me a real world please with big game and program selection.

The 1600 vs 7600k was also a joke at those resolutions two years ago. Now the 1600 is the better CPU overall, outside of Arma, FarCry, and Tomb Raider.

A lot of reviewers tend to benchmark CPU's in games looking at a wall or the sky, where they get the most FPS because somehow they equate that to CPU performance.

In fact it's not, where there's nothing going on in the scene the CPU isn't actually being used, at least not properly, it's how they come up with really idiotic conclusions like this higher clocked 4 core is better than this 8 core, yet if they were to look away from the wall or the sky that 4 core craters as all of its poultry 4 thread's are maxed out while the 6 shoots into the lead.

Digital Foundry did some great 1600 vs 7600K testing where they demonstrated this, with nothing to see in the scene the 7600K was pushing 240 FPS vs 210 on the 1600, literally looking at the sky, looking at the complex vista at ground level the 7600K tanked to 60 FPS while the 1600 was at 110 FPS, nearly twice as fast. so which CPU is actually better?

About 90% of reviewers said the 7600K was better, that tells you everything you need to know about those reviewers, there all idiots.

 

[DragonQ]

Power usage is interesting.

3700X using way more power than the 2700X in non gaming and the 3900X comparable to the 9900k

Edit: From the same article on Reddit

Likely the new AVX instructions boosting power usage in non-gaming.

 

[Deleted User 456458]

At this stage, why would you expect the 3900x to have a gaming advantage over the 3700x? How many game engines are designed for 12 core usage (or even 8?). I don't understand where some of this comes from.

 

[DragonQ]

At this stage, why would you expect the 3900x to have a gaming advantage over the 3700x? How many game engines are designed for 12 core usage (or even 8?). I don't understand where some of this comes from.

I think some thought that the higher boost clocks would help but even if it did it was always going to be a tiny percentage increase.

 

[nashathedog]

Just remember there are plenty X470 boards that will handle an overclocked 12 core and even 16 core. I'll be buying X470 for sure if the X570 prices are as daft as it appears.

Spoiler

It looks like I lucked out as I went with the Asus Prime X470 Pro for my 2700x which has all green except an orange box for overclocking the 3950x, I'll go for a 3900x, I would like to move onto pcie-4 for the nvme drives but dropping a 3900x in to replace the 2700x is okay, especially if X570 pricing's sky high.

 

[Deleted User 456458]

I think some thought that the higher boost clocks would help but even if it did it was always going to be a tiny percentage increase.

We'll have to see what the actual sustainedd boost is better determine that. Also offset it against any multi chip performance overhead.

 

[humbug]

At this stage, why would you expect the 3900x to have a gaming advantage over the 3700x? How many game engines are designed for 12 core usage (or even 8?). I don't understand where some of this comes from.

4.6Ghz vs 4.4Ghz boost clocks.

And it should, there's something a little odd about it.

 

[4K8KW10]

4.6Ghz vs 4.4Ghz boost clocks.

And it should, there's something a little odd about it.

Something odd in double the GameCache size

 

[humbug]

Something odd in double the GameCache size

X2 35MB, one for each Die.

 

[kitfit1]

And it should, there's something a little odd about it.

Is there ? i don't think so. We are seeing early results from very early sku's. Not only that, i don't think for one min these are the only sku's we will see. We already know the will be a 3950x, i bet there will be a 3650x a 3750x and a 3850x at some point................................maybe even at the launch of the 3950x.

 

[humbug]

Is there ? i don't think so. We are seeing early results from very early sku's. Not only that, i don't think for one min these are the only sku's we will see. We already know the will be a 3950x, i bet there will be a 3650x a 3750x and a 3850x at some point................................maybe even at the launch of the 3950x.

Well that could explain the oddity.

 

[Deleted User 456458]

Interesting to see which games engines are barely multi core and really want clock speed. When you see the 7700k ahead of the 3xxx it's the legacy nature of the engine. If you look at Origins benchmark, you can see how the 7700k falls to the bottom.

As time goes by and next gen of consoles force multicore engines across the board, the 3900x will start coming into it's own. However, if your upgrade cycles is 2 years or, who cares and buy what's fastest today.

There you go. You can sleep in Sunday now.

 

[LordByron69]

Everyone said the same when Ryzen launched, and kept recommending the 7600k over the 1600. Now the latter is overall better in games also.

I find 480 and 720 testing pointless, it’s about as synthetic as it gets. As it doesn’t account for new game engines, more multi threading support, and developers developing on and with Ryzen to optimize for the architecture.

It’s clear that older, and some specific games still struggle with Ryzen, such as ancient Arma, and FarCry which doesn’t seem to run well on the AMD architecture at all, even if the IPC and clocks are identical. Those are outliers; which is why we need more than just 5-6 games tested at a resolution that 16% of all steam users even run ( most likely low end systems and laptops ).

I said CPU bottleneck scenario.
My point was not about benchmarking an empty sky...

 

[DylRicho]

Even without playing at this low resolution, it's important to benchmark in CPU bottleneck scenario to estimate the future proof potential in few years when games will be more demanding on CPU even at higher resolution.

I'm not convinced that these two scenarios affect the CPU in the same exact way. Nor do I believe it to be a worthwhile test.

Five years from now, the most popular resolution is still likely to be 1080p, and I believe so for two reasons:

1) People just seem to like 1080p as the perfect balance between real estate and performance compromise. 900p is also a good candidate for this (I know that some people prefer to not play in full-screen mode, so as to make multi-tasking easier). 1080p panels are also the most common sold.

2) APUs. Ryzen APUs have practically killed the low-end dedicated GPU segment. You can now enjoy GT 1030 levels of performance in a $99 APU. Go back five years to 2013. The HD 8670D (A10-6800K) was the integrated graphics king. Using the same game (and therefore game engine), which is Grand Theft Auto V:

• HD 8670D (A10-6800K) = 720p / normal / 40 fps avg
• Vega 8 (Ryzen 3 2200G) = 1080p / normal / 60 fps avg

With an overclock, you can bump up the settings to a medium-high balance and not lose frames. In five years, you've gained 50% more frames and increased the number of on-screen pixels by 125%. With the 2200G as the baseline, think what can be achieved in another five years. APUs will take on some serious improvements in the years to come, most likely to the point where AMD is actively pushing 1440p/60 fps marketing material.

If we're supposed to take these Zen 2 720p benchmarks as some kind of insight into the future, I think it's incredibly misleading.

Intel sponsored the 720p testing.

Would this be considered one of their "real-world" tests?

According to Intel, one of the benchmarks we should be testing, is how quickly Skype opens.

And SYSmark, of course. It would be illegal to forget about that.