AMD Ryzen (Summit Ridge) Benchmarks Thread (use new thread)

[itsmydamnation]

every time i say the word ryzen to myself i end up here in my head:

it just matches so well

 

[krumme]

It's ramping production. It will obviously be a higher bin. These samples come from small batches with very little opportunity to harvest higher clocks parts. Also there are things like steppings and process tweaks that can increase clocks.

Lisa seamed confident. She said "base speeds of 3.4Ghz and higher" So there will be parts with higher base clocks than 3.4Ghz.

Notice 3.4 or higher. Not "and".

 

[Arachnotronic]

every time i say the word ryzen to myself i end up here in my head:

it just matches so well

This is a really good song. Thanks for sharing!

 

[itsmydamnation]

This is a really good song. Thanks for sharing!

I only listen to the most awesomest of metal

 

[Arachnotronic]

I only listen to the most awesomest of metal

You should post more metal samples, your taste seems excellent. Following this group on Spotify now

 

[rvborgh]

This looks really quite promising. Certainly far better than anything AMD has come out with in the past and most importantly... it puts them back into the running to make some decent money. Who says that AMD isn't sandbagging the frequency? If this is a first showing... its a great one.

i'm quite happily looking forward to building at least two new PCs quite soon with these

 

[itsmydamnation]

my only concern now is the memory controller how high can we clock them DDR4's or if we cant clock them that high how low can we get timings.

 

[itsmydamnation]

You should post more metal samples, your taste seems excellent. Following this group on Spotify now

You will probably want to check out Matt Barlow era ICE EARTH (same vocalist as that pyramaze album), then if you like that, Demons & wizards ( same guitarist as ice earth (they also play my most fav song of all time)) and then finally my favorite blind guardian ( same vocalist as Demons and wizards).

All i need is some one from blind guardian to do a pyramaze album and then the circle will be complete...lol

 

[sm625]

You will probably want to check out Matt Barlow era ICE EARTH (same vocalist as that pyramaze album), then if you like that, Demons & wizards ( same guitarist as ice earth (they also play my most fav song of all time)) and then finally my favorite blind guardian ( same vocalist as Demons and wizards).

All i need is some one from blind guardian to do a pyramaze album and then the circle will be complete...lol

lol I just listened to something wicked yesterday... I came to like avenged sevenfold in large part because they remind me a lot of iced earth, but fortunately with a much better singer.

At any rate I wonder if the 3.4GHz sku will be very highly binned with only very limited supply.

 

[sirmo]

every time i say the word ryzen to myself i end up here in my head:

it just matches so well

Similar, this is what I was thinking of:

 

[itsmydamnation]

At any rate I wonder if the 3.4GHz sku will be very highly binned with only very limited supply.

so lets make it clear

$zen_schedule = launch;
$base_freq = 3400 + int rand();
$conditions_met=0;

until ( $conditions_met ==1){
    if ($zen_schedule eq 'launch'){
        (if $base_freq == 3400 || $base_freq > 3400){
            $conditions_met=1;
        }
    }
}

lisa says "our horizon processors at launch will have base clock speeds of 3400mhz or higher";

 

[Thunder 57]

so lets make it clear

$zen_schedule = launch;
$base_freq = 3400 + int rand();
$conditions_met=0;

until ( $conditions_met ==1){
    if ($zen_schedule eq 'launch'){
        (if $base_freq == 3400 || $base_freq > 3400){
            $conditions_met=1;
        }
    }
}

lisa says "our horizon processors at launch will have base clock speeds of 3400mhz or higher";

Does rand() return a signed or unsigned int? I kid, I kid

 

[Arachnotronic]

A hint on Ryzen pricing:

This guy implies that top Ryzen will be cheaper than 6900K.

 

[sirmo]

It's a pretty much a given it will be cheaper. AMD needs to break the inertia of people buying Intel. Also there is plenty of room for AMD to price the Ryzen for decent margins. There will probably be a top bin gold sample part that's going to be expensive. But I am convinced the main SKUs will be significantly cheaper than Intel.

If AMD was going after the $1k CPU market they would have included quad channel memory. Their intentions are clear, 8c/16t will be sold at mainstream prices. Around $500 or less.

 

[Arachnotronic]

It's a pretty much a given it will be cheaper. AMD needs to break the inertia of people buying Intel. Also there is plenty of room for AMD to price the Ryzen for decent margins. There will probably be a top bin gold sample part that's going to be expensive. But I am convinced the main SKUs will be significantly cheaper than Intel.

If AMD was going after the $1k CPU market they would have included quad channel memory. Their intentions are clear, 8c/16t will be sold at mainstream prices. Around $500 or less.

The lack of quad channel memory is reflective of AMD's strategy in servers. Sever chips are MCMs of 8 core/2 channel IMC dies. Don't think that enthusiast desktops influenced that decision

 

[Doom2pro]

I think we can all agree the 6900K is overpriced, vastly... Don't even get started on the 6950k... AMD obviously isn't going to join Intel in over pricing their CPUs that perform as good as if not better than the 6900K. Intel is only over pricing these products due to lack of competition from AMD.

AMD's top offerings should be significantly less than Intel's 6900K MSRP, and they will draw in a lot of people who want that level of performance but not at the Intel price.

I think it's safe to assume there is a large market of people out there who would be willing to spend a little more to buy a $400-$500 Ryzen that performs as much as an Intel CPU they would have never considered buying. I have a feeling Ryzen is going to sell like hotcakes in a famine during a blizzard.

 

[sm625]

Does rand() return a signed or unsigned int? I kid, I kid

rand() returns a signed int.

 

[sm625]

The blue graph pretty clearly shows Intel cpus are only gaining max 1.5% performance per 100 Mhz increment. There is no even remotely reasonable clock speed that is going to propel Ryzen anywhere near 6700k territory. Even at 5 GHz it will still lose to a stock 6700k, unless the cpu has otherworldly scaling.

 

[sirmo]

The blue graph pretty clearly shows Intel cpus are only gaining max 1.5% performance per 100 Mhz increment. There is no even remotely reasonable clock speed that is going to propel Ryzen anywhere near 6700k territory. Even at 5 GHz it will still lose to a stock 6700k, unless the cpu has otherworldly scaling.

The scaling isn't linear though. At low 3Ghz it could be 2-3% per 100Mhz.. and only .5% at 4Ghz. Which makes sense, because once you remove the CPU bottleneck the GPU becomes one. And most games are optimized for some baseline like 3.7Ghz Intel CPU.

Compared to higher clocked i5s Ryzen looks really solid despite its 3.15Ghz base and 3.3Ghz boost. All these games tend to not scale past 4 cores. In either case it looks very competitive to me.

 

[PhonakV30]

It's a pretty much a given it will be cheaper. AMD needs to break the inertia of people buying Intel. Also there is plenty of room for AMD to price the Ryzen for decent margins. There will probably be a top bin gold sample part that's going to be expensive. But I am convinced the main SKUs will be significantly cheaper than Intel.

If AMD was going after the $1k CPU market they would have included quad channel memory. Their intentions are clear, 8c/16t will be sold at mainstream prices. Around $500 or less.

High Speed of DDR4 rather than Quad channel.Hope AMD Zen offers up to DDR4 3200

 

[Arachnotronic]

Compared to higher clocked i5s Ryzen looks really solid despite its 3.15Ghz base and 3.3Ghz boost. All these games tend to not scale past 4 cores. In either case it looks very competitive to me.

Let's fact check this

The games tested are BF4, ARMA III, X3: TC, Witcher 3: Wild Hunt, Anno 2070, GRID: Autosport, and Far Cry 4.

How do these games all scale with cores?

Far Cry 4

This game seems to like fewer, but faster cores. Also doesn't seem to respond much to architectural improvements (no difference between 3770K and 4770K, for example). Mainly seems to like clock speed (as the 4960X's result shows, a chip that's 3.6GHz base/4GHz turbo).

Battlefield 4

This game seems to run well on everything, from FX 8350 to 4960X. I don't think this is a good measure of IPC for obvious reasons.

Anno 2070

Hard to tell the difference between the CPUs at 1080p, so GPU limited. Even at 1024x768, it seems that scaling past 4 threads is poor and once again it doesn't respond all that well to architectural improvements (basically no difference between SNB/IVB/HSW at same clocks).

Again, doesn't look like a good IPC test since it doesn't really respond to CPUs with measurably better IPC.

ARMA 3

Doesn't seem to scale well past four threads (3570K vs 3770K result), nor does it seem to be all that sensitive to core perf/MHz (look at how well the FX 8350 is holding up and we know that's a much weaker core than IVB/SNB). Even i7 920 is holding its own despite a big frequency/IPC disadvantage to the newer core.

Once again, doesn't look like a game that's all that sensitive to better microarchitecture/higher IPC.

GRID Autosport
This game seems to respond reasonably well to more cores up through about five cores/five threads. Then it just doesn't seem to care. Big jump in going from 2 threads to five threads, though.

http://www.dsogaming.com/pc-performance-analyses/grid-autosport-pc-performance-analysis/

X3: Terran Conflict

This game seems to like IPC/clocks more than anything else, doesn't scale too well with core/thread count (see: 2600K vs 2500K, and 2600K vs 3960X).

Witcher 3: Wild Hunt

This game doesn't seem super sensitive to CPU performance, FX 8350 does almost as well as a 2500K, which does nearly as well as a 3770K.

Doesn't seem like a good test of CPU IPC/performance.

Anyway, so looks like we have a bunch of tests that, in aggregate, probably don't tell us much about the underlying CPU core performance.

We can't really conclude much about relative IPC (single thread perf/MHz) of the various architectures based on this set of games, IMO.

 

[majord]

Let's fact check this

The games tested are BF4, ARMA III, X3: TC, Witcher 3: Wild Hunt, Anno 2070, GRID: Autosport, and Far Cry 4.

How do these games all scale with cores?

Far Cry 4

This game seems to like fewer, but faster cores. Also doesn't seem to respond much to architectural improvements (no difference between 3770K and 4770K, for example). Mainly seems to like clock speed (as the 4960X's result shows, a chip that's 3.6GHz base/4GHz turbo).

Battlefield 4

This game seems to run well on everything, from FX 8350 to 4960X. I don't think this is a good measure of IPC for obvious reasons.

Anno 2070

Hard to tell the difference between the CPUs at 1080p, so GPU limited. Even at 1024x768, it seems that scaling past 4 threads is poor and once again it doesn't respond all that well to architectural improvements (basically no difference between SNB/IVB/HSW at same clocks).

Again, doesn't look like a good IPC test since it doesn't really respond to CPUs with measurably better IPC.

ARMA 3

Doesn't seem to scale well past four threads (3570K vs 3770K result), nor does it seem to be all that sensitive to core perf/MHz (look at how well the FX 8350 is holding up and we know that's a much weaker core than IVB/SNB). Even i7 920 is holding its own despite a big frequency/IPC disadvantage to the newer core.

Once again, doesn't look like a game that's all that sensitive to better microarchitecture/higher IPC.

GRID Autosport
This game seems to respond reasonably well to more cores up through about five cores/five threads. Then it just doesn't seem to care. Big jump in going from 2 threads to five threads, though.

http://www.dsogaming.com/pc-performance-analyses/grid-autosport-pc-performance-analysis/

X3: Terran Conflict

This game seems to like IPC/clocks more than anything else, doesn't scale too well with core/thread count (see: 2600K vs 2500K, and 2600K vs 3960X).

Witcher 3: Wild Hunt

This game doesn't seem super sensitive to CPU performance, FX 8350 does almost as well as a 2500K, which does nearly as well as a 3770K.

Doesn't seem like a good test of CPU IPC/performance.

Anyway, so looks like we have a bunch of tests that, in aggregate, probably don't tell us much about the underlying CPU core performance.

We can't really conclude much about relative IPC (single thread perf/MHz) of the various architectures based on this set of games, IMO.

Most of what you posted were GPU limited tests, We don't even know the details on how the games were tested by CanadrdPC, and which were CPU limited, etc, except to say there's clear scaling between the i5's with different frequencies, so, on average there is at least some CPU limitation.

Given most of the titles tested are quite old, we know by default they won't scale well with higher thread count, and it's plain obvious from the fact Skylake is soundly beating BW-E. Something that rarely happens in modern titles.

 

[jelome1989]

Kinda offtopic, but just want to add, if we're talking about games, IMO, Starcraft 2 would be the most revealing game if you want to test a CPU's single core gaming performance. Starcraft 2 should be in every single game-based CPU benchmarks. Not to mention since the game is absolutely CPU limited, memory speeds also scale well with OCs.

 

[krumme]

@Dresdenboy and the likes. What about a zen+ arch thread?
We want crazy stuff like 5 wide, smt 4t/core, no avx2

 

[guskline]

Waiting for RyZen to officially lauch feels like being a child on Christmas Eve!

PLEASE SANTA when are you going to get here!