Leaked AMD Ryzen Benchmarks?

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
The French website CPC Hardware has posted what they claim to be AMD Ryzen performance benchmarks. If these numbers are to be believed, the Ryzen's performance looks pretty damn good considering this is an engineering sample in an unknown motherboard running games and benchmarks that can't possibly be optimized for the CPU - motherboard combo.
 
I question it.

Makes no sense that some french site would have data before anyone else.


That, and they are "aggregate" results, so they are pretty much meaningless. Can't be extrapolated in any useful way.

If it is real, it looks pretty bad IMHO. It may be in the top three on the computational chart (green), but that's with double the number of cores compared to many of the other chips in that lineup.

Per core performance is still king, and it if can't deliver there, it will be a also-ran for our purposes.

As one might expect from a chip that suffers in the per core department, it does pretty poorly in the game benchmark (blue)

Time will tell though. I'm not writing anything off yet.
 
I question it.

Makes no sense that some french site would have data before anyone else.


That, and they are "aggregate" results, so they are pretty much meaningless. Can't be extrapolated in any useful way.

If it is real, it looks pretty bad IMHO. It may be in the top three on the computational chart (green), but that's with double the number of cores compared to many of the other chips in that lineup.

Per core performance is still king, and it if can't deliver there, it will be a also-ran for our purposes.

As one might expect from a chip that suffers in the per core department, it does pretty poorly in the game benchmark (blue)

Time will tell though. I'm not writing anything off yet.
Somebody always has data before someone else, so why not the French. The chip is probably made down the road in Dresden, though I don't know where it's packaged, but even that does not mean anything.
It sure does not look like they faked the power usage. ;)
Time will tell.
 
If some of you really truly honestly think the AMD Ryzon results are going to be substantially different than these results, I would say this to you. Manage your expectations better.

ES sample chip this, beta motherboard that. AMD cannot beat Intel on single threaded performance but for productivity, they can come close or slightly ahead. If they get the boost speed up higher, they can come close to the 6700K single threaded performance.

As the Ryzon stands now game wise, it's below Intel's i5 and close to i3 single threaded performance. according to some of this material that was leaked.

Price is really a big concern for a lot of people. The general feeling is that people would consider the Ryzon, even at lesser performance than Intel if the pricing was around $350 or so. If not it's just going to be smarter from a pricing and performance POV to buy Intel.

AMD knows this and I am sure they have their marketing people all over this small detail.

You know what the difference is between Intel and AMD chip development? Several billion dollars. AMD was never going to release a single-threaded Intel killer. The one comment that stands out for me during all of this is someone mentioned that AMD has managed to catch up to Intels 2011.

If they get the boost speeds up or you can overclock Ryzon to 4ghz+ and it costs arund $350, then I'll upgrade to AMD, no doubt.
 
Last edited:
If some of you really truly honestly think the AMD Ryzon results are going to be substantially different thank these results, I would say this to you. Manage your expectations better.

ES sample chip this, beta motherboard that. AMD cannot bet Intel on single threaded performance but for productivity, they can come close or slightly ahead. If they get the boost speed up higher, they can come close to the 6700K single threaded performance.

As the Ryzon stands now game wise, it's below Intel's i5 and close to i3 single threaded performance. according to some of this material that was leaked..

it's not far off an I5 really, it's clocked low in these benchmarks, I don't think performance increases more than 2-3 % from an ES, having playd with a lot of es's, and in most scenarios it does not have any ipc difference than the real deal - Just consumes more power, no fancy boost or frequency control.
However, clocked higher it'll perform like a 2 gen old intel and in multithread on par.
 
I only went Intel a few years back because of the platform tech being so much better than AMDs at the time. Hell, I still feel dirty about it. :D

If AMD can deliver there plus CPU performance, I'd drop Intel in a heartbeat.
 
I question it.

Makes no sense that some french site would have data before anyone else.


That, and they are "aggregate" results, so they are pretty much meaningless. Can't be extrapolated in any useful way.

If it is real, it looks pretty bad IMHO. It may be in the top three on the computational chart (green), but that's with double the number of cores compared to many of the other chips in that lineup.

Per core performance is still king, and it if can't deliver there, it will be a also-ran for our purposes.

As one might expect from a chip that suffers in the per core department, it does pretty poorly in the game benchmark (blue)

Time will tell though. I'm not writing anything off yet.

You kinda make no sense. Most games are optimized for up to four cores except for some. It shines in multi core cuz it has ipc to back it up, makes sense? It's slower in games because clock speed is only 3.15ghz on this samples. Heck even the 6900k struggles because you see it doesn't have the clock speed matching the 6700k but in games that can use all the cores they will both be able to sustain bettter minimum frame rates.

So let's say in games only using 4 cores may be zen's auto overclocking feature can get up to like 3.8 or so. We just don't know all we know is zen will be 3.4ghz base clock minimum for the retail chips. No other numbers were given.
 
WoW, that's a week old. Anyway, here's the translation.

[] Performances - Raw computations

H.264 encoding 1080p & H.265, Wprime, PovRay 3.7, Blender 3D, 3DSMax 2015 / Mental Ray, Corona Benchmark

With it's true eight cores, Zen accomplishes prowesses despite it's limited frequency of 3.3 Ghz. It's getting dangerously close - for Intel - of the Intel 6900K while offering performance comparable to the Core i7 5960X which has an identical frequency (3.3 Ghz in Turbo mode). AMD's allegations a few months ago seem to check out in practice, and this is an excellent news. Compared to the FX-8370, we witness a performance gain of around 35% for equal clockspeeds, matching there too with the manufacturer's previsions (40%)

[] Video games performances

Far Cry 4, GRID: AutoSport, Battlefield 4, Arma III, X3:TC, The Witcher 3 : Wild hunt, Anno 2070

If the results might seem clearly more disappointing on the average of the tested games, it fits to keep in mind than the tested prototype was an octo-core with a quite low frequency (in particular in Turbo mode). Meanwhile, the tested games stay very sensitive to the frequency and keep struggling to exploit more than four cores. Hard in those conditions to compare it with a Core i7 6700k of which frequency is at over 4 Ghz. Remains that the Zen architecture shows an efficiency we haven't saw at AMD since a very long time.

[] Electric consumtion

Measures took at full load (in watts)

The measurement of the electric consumption of the Zen CPU was took with an amps meter on the ATX 12V connector at full load. While this method is less precise than the usual one we do on the oscilloscope, it gives a good idea of the performances of the 14 nm LPP of Global Foundries. Once removed losses from the motherboard's VRMs, we can estimate that the processor consumes a bit less than 90 W, a value very close to that of an 6900K. A result that gives good faith in the future.

Editor's opinion:

The Zen architecture should allow ZEN to seriously come back on the CPU market, including the high-end segment abandoned since long. Intel thus risks to suffer a violent [untranslatable French gibberish], a well-deserved consequence for their years of lazyness and arrogance ( The 1 900€ core i7 6950X shall stay the perfect illustration of that). If we hope the comeback of a real concurrency on the processor market, all is not won yet for AMD. While the octocores look on schedule, the constructor must finalize quickly the quad-core derivatives with vastly higher frequencies than the current prototypes: 3.8 see 4.0 or 4.2 Ghz seems like the minimum to tickle the latest Kaby Lake. At last, remains a sizeable unknown : the prices. If some rumours talk of a prety low price grid, we doubt AMD will cut down their CPU while they at last have a chance to make up for their years of losses. At must, do not fuck up in the last straight line ...
 
I question it.

Makes no sense that some french site would have data before anyone else.


That, and they are "aggregate" results, so they are pretty much meaningless. Can't be extrapolated in any useful way.

If it is real, it looks pretty bad IMHO. It may be in the top three on the computational chart (green), but that's with double the number of cores compared to many of the other chips in that lineup.

Per core performance is still king, and it if can't deliver there, it will be a also-ran for our purposes.

As one might expect from a chip that suffers in the per core department, it does pretty poorly in the game benchmark (blue)

Time will tell though. I'm not writing anything off yet.

The magazine's chief editor is Samuel Demeulemeester. I saw on Anandtech's forum this post, and so did he apparently. Make of it what you will but the only thing I've seen is potentially french posters say they are legit but i dunno myself.
 
As the Ryzon stands now game wise, it's below Intel's i5 and close to i3 single threaded performance. according to some of this material that was leaked.
Are you aware that single threaded performance is pretty much the same on the i3, i5 and i7 at the same clockspeed?
 
The French website CPC Hardware has posted what they claim to be AMD Ryzen performance benchmarks. If these numbers are to be believed, the Ryzen's performance looks pretty damn good considering this is an engineering sample in an unknown motherboard running games and benchmarks that can't possibly be optimized for the CPU - motherboard combo.

The person who did the benchmarks is a bit more reliable than some random dude on the internet though. :) However, even he said to take things not as gospel since it was not a final release and it is all on beta hardware. We will see though, we will see.
 
If a game uses one thread now days it is a crappy poorly optimize game that will never use any processor capability on the planet. For older games Zen will most probably be more then sufficient if you are into old games.

I could care less about one thread or two threaded games performance - they are most likely not AAA games or very bad ones. DX 12 and Vulkan is the future so more cores used, cpu becomes less a limitation. Also if the minimums are above your monitor refresh rate or is seldom below it - it doesn't matter if 1000fps min or 60fps in worst case scenarios if your monitor is 60hz anyways - well at least for me that is. Gaming experience will not be affected. Of course argument that the faster one will have a longer life span could be used.

If you have programs that uses all the cores available then going for a more core chip with decent IPC will most likely be the winner if the price is right. Let see - IPC of Skylake I3 = I5 = I7 or should I say is close due to cache differences. I don't think anyone expects Zen IPC to = Skylake per core but it may multithread better giving it about equal performance for same core count and clock speed of Intel for a hell a lot better price (What I am really looking at here).
 
If final hardware and Ryzen 8c/16t performs at Haswell-E levels, I'm expecting 650$ and I would say it's worth it. Again just my opinion. We can't expect AMD to price themselves out of existence.
 
If it can come colse to Intel it would be a win as then intel might release a cpu that gets more then a 5% bump per gen.
Now if this thing can come colse and oc to mid 4s or higher and the rest of the platform is upto par then it will be a super win for amd. Sadly after how lacking amd has been all around im not holding my breath
 
If final hardware and Ryzen 8c/16t performs at Haswell-E levels, I'm expecting 650$ and I would say it's worth it. Again just my opinion. We can't expect AMD to price themselves out of existence.
I agree with that, a lower price if the motherboards are not as strong comes into play as well. It is the whole eco system that has to come through and not only the processor. It needs to support DDR 4 memory as fast as possible, especially being 8 cores and only a dual channel processor. All modern and previous features well supported on the motherboard. A very complex launch and something AMD has not done in awhile - so I say be very aware of everything. If AMD and AIBs for the motherboards hit it out of the park then I say buy Zen soon, if not I say wait for Skylake E and then do another comparison.
 
Since Ryzen has only a 95W TDP we won't need extra powerful cooling or an overly expensive AIO to get the highest clock speeds as it was mentioned at New Horizon. I'm just really stoked. :)
 
I agree with that, a lower price if the motherboards are not as strong comes into play as well. It is the whole eco system that has to come through and not only the processor. It needs to support DDR 4 memory as fast as possible, especially being 8 cores and only a dual channel processor. All modern and previous features well supported on the motherboard. A very complex launch and something AMD has not done in awhile - so I say be very aware of everything. If AMD and AIBs for the motherboards hit it out of the park then I say buy Zen soon, if not I say wait for Skylake E and then do another comparison.

Good call.
 
I am pretty stoked as well, although I am still using my 2700k with my 1080 card, and see no reason to upgrade yet. I am just happy to see AMD playing ball again, in the same field. It'll bring prices down, and maybe I might end up with a reasonably decent setup for cheap again should this MB die or such. I already had to replace it twice. This CPU still goes 4.4 all day rock solid. I'm not going to pay $350 for a ~5% performance upgrade. Unless it breaks, but I want this to last as long as it takes for AMD to be a player again, just to save my pocketbook.
 
I don't blame you one bit on that at all. I'm on an A10-7850K heavily oc'd and while I'm enjoying it for the most part I can't play Doom at an fps that's satisfactory for me. 22fps isn't going to cut it. lol
 
If some of you really truly honestly think the AMD Ryzon results are going to be substantially different than these results, I would say this to you. Manage your expectations better.

ES sample chip this, beta motherboard that. AMD cannot bet Intel on single threaded performance but for productivity, they can come close or slightly ahead. If they get the boost speed up higher, they can come close to the 6700K single threaded performance.

As the Ryzon stands now game wise, it's below Intel's i5 and close to i3 single threaded performance. according to some of this material that was leaked.

Price is really a big concern for a lot of people. The general feeling is that people would consider the Ryzon, even at lesser performance than Intel if the pricing was around $350 or so. If not it's just going to be smarter from a pricing and performance POV to buy Intel.

AMD knows this and I am sure they have their marketing people all over this small detail.

You know what the difference is between Intel and AMD chip development? Several billion dollars. AMD was never going to release a single-threaded Intel killer. The one comment that stands out for me during all of this is someone mentioned that AMD has managed to catch up to Intels 2011.

If they get the boost speeds up or you can overclock Ryzon to 4ghz+ and it costs arund $350, then I'll upgrade to AMD, no doubt.



If the internet was as vast then as it was now, I'd imagine someone would have said the same thing about AMD's Athlon back in the 90's, saying AMD just made 3rd rate chips like Cyrix and have heat issues, right before the Athlon burst onto the scene and kicked Intel's ass for 3 straight years.

The thing is, while everyone likes to think themselves an expert, less money spent smartly will almost always win over a ton of money spent poorly - and when there is no competition to keep a company sharp, they tend to let their guard down - it's how AMD did it the first time.


Not saying history will repeat itself, but not betting against it like you are either.
 
R&D money is not a good indicator for success nor lack of it. Intel spent a lot of money into the process tech which AMD does not have to do is another consideration.

AMD gets the benefits of Glo-Flo and SamSung as well as TSMC R&D for process tech and gets to pick and choose anyone or all of them as they see fit.

A good team that clicks and has access to a broad menu of items they can use can succeed in making something that would be considered impossible previously. So maybe we will see some real excitement come back to the PC cpu area for once.
 
If some of you really truly honestly think the AMD Ryzon results are going to be substantially different than these results, I would say this to you. Manage your expectations better.

ES sample chip this, beta motherboard that. AMD cannot bet Intel on single threaded performance but for productivity, they can come close or slightly ahead. If they get the boost speed up higher, they can come close to the 6700K single threaded performance.

As the Ryzon stands now game wise, it's below Intel's i5 and close to i3 single threaded performance. according to some of this material that was leaked.

Price is really a big concern for a lot of people. The general feeling is that people would consider the Ryzon, even at lesser performance than Intel if the pricing was around $350 or so. If not it's just going to be smarter from a pricing and performance POV to buy Intel.

AMD knows this and I am sure they have their marketing people all over this small detail.

You know what the difference is between Intel and AMD chip development? Several billion dollars. AMD was never going to release a single-threaded Intel killer. The one comment that stands out for me during all of this is someone mentioned that AMD has managed to catch up to Intels 2011.

If they get the boost speeds up or you can overclock Ryzon to 4ghz+ and it costs arund $350, then I'll upgrade to AMD, no doubt.
I think you need to know what you're talking about better.

If you've paid attention to the market for the past 10 years you'd realize that more cores in one package = lower frequency to keep TDP reasonable.

Example:
4c/8t: 4GHz
6c/12t: 3.6GHz
8c/16t: 3.2GHz

Of course for applications that are not heavily threaded (gaming) and are more clockspeed/IPC dependent, fewer and faster cores is better than too many slower cores. The Zen core is matching perf from comparable Intel products and that's all you need to know, really. When AMD reveals 4c/6c products expect the base clocks to be more comparable to Intel's 4c and 6c offerings as well since TDP headroom will allow it.

As was previously said, i3/5/7 single core IPC is virtually identical, less relatively small differences due to cache differences.

Be smarter when trying to discredit AMD, man
 
I question it.

Makes no sense that some french site would have data before anyone else.

These are the SAME guys who leaked the original Athlon 64 way back in early 2000 months before it was launched. This tech magazine has a long history and a lot of connections in the industry.

Also the aggregate benchmarks for the gaming results makes perfect sense if Ryzen ~ Broadwell-E in IPC since the sample is a 3.15ghz with 3.3ghz boost. Many of those games they tested are quite old and do not take advantage of 4+ cores/threads and it's against Broadwell-E that turbo to 4ghz on low-thread workloads.

The results show Ryzen is definitely much faster than the FX architecture, how fast? ~Broadwell-E and even Skylake for IPC.

If true and priced well, AMD & PC enthusiasts are going to have a hell of a great 2017.
 
Until retail copies are available for general consumption, I'm going to take all benchmark claims with a couple dozen brine blocks.
 
Worst case scenario, I'll be able to finally buy myself an unlocked i3, chill it and overclock it to 11.

They ought to go down in price, too. Right now a used i3-6300 goes for around 100 dollars.

Best case scenario - I'll get to relive my golden days of abusing my s754 and s939 Athlon 64.

I'm pretty happy at this point. Now, the money part. :D
 
I believe them. It sorta feels like the launch of bulldozer all over again, but a little better. Its 24% faster then a FX8370 at gaming but with a 25% clock speed deficit. Im currently gaming on a stock clocked FX8320e so it still sounds like a huge upgrade for me (that i also do not need)
 
When AMD reveals 4c/6c products expect the base clocks to be more comparable to Intel's 4c and 6c offerings as well since TDP headroom will allow it.

That's what I'm more waiting for

Not even half interested in a 8 core part

But if AMD could bring out a 6 core at ok clocks (~4Ghz, same TDP should translate to higher clocks on less cores, maybe some boosting headroom under an AiO) at 350$-380$ they'd already have a winner
A 6800k is 450$
Undercutting Intel by 100 bucks for close enough performance would be quite something
 
I looked over everything like most here. Some of the numbers that stood out to me was gaming performance.
I think you need to know what you're talking about better.

If you've paid attention to the market for the past 10 years you'd realize that more cores in one package = lower frequency to keep TDP reasonable.

Example:
4c/8t: 4GHz
6c/12t: 3.6GHz
8c/16t: 3.2GHz

Of course for applications that are not heavily threaded (gaming) and are more clockspeed/IPC dependent, fewer and faster cores is better than too many slower cores. The Zen core is matching perf from comparable Intel products and that's all you need to know, really. When AMD reveals 4c/6c products expect the base clocks to be more comparable to Intel's 4c and 6c offerings as well since TDP headroom will allow it.

As was previously said, i3/5/7 single core IPC is virtually identical, less relatively small differences due to cache differences.

Be smarter when trying to discredit AMD, man


We will see soon enough.

I will say this. There is an awfully lot of smoke and mirrors surrounding Ryzon. And where there's smoke there 's .... never mind, you know the rest.
 
It appears as fast as comparably clocked Intel parts in gaming as it beats the i5 at 3.2/3.5 despite being clocked a little lower (ignoring core count which most games don't utilize above 4 threads), which is very good news as long as they can clock it comparably. We already know the 7th gen parts don't improve ipc at all for Intel and apparently don't improve clocks much either, so this should be a pretty straight fight at launch.

It will be interesting to see what the 4c parts can do clock wise. That might be the hot ticket for a full platform switch and then move to the 8 core part on the 3rd rev when hopefully more stuff is finally threaded enough to make use of them. The nice thing is you won't have to switch platforms at that point (unlike Intel which will probably have at least one socket switch in that time).
 
Example:
4c/8t: 4GHz
6c/12t: 3.6GHz
8c/16t: 3.2GHz

Ye that's not going to happen. You only have to look at current products to understand that.

Also you forget what the architecture and/or process may be limited to.
 
Ye that's not going to happen. You only have to look at current products to understand that.

Also you forget what the architecture and/or process may be limited to.
he was showing how the clock reduces as core/thread goes up. hes not guessing the ryzen speeds.
 
he was showing how the clock reduces as core/thread goes up. hes not guessing the ryzen speeds.

It still depends on the architecture and process node. Its not some kind of guarantee that it will happen.

Despite SB being a great overclocker, 6 cores came in at 3.3/3.9 Ghz and quad in at 3.4/3.8 or 3.5/3.9 depending on how you see it. With IB you ended up with 6 cores at 3.6/4.0Ghz while the quad managed 3.5/3.9Ghz.
 
Last edited:
If a game uses one thread now days it is a crappy poorly optimize game that will never use any processor capability on the planet. For older games Zen will most probably be more then sufficient if you are into old games....

It's not as simple as that, there are just a limited number of tasks within a game that benefits from multiple cores, graphics are already taken care of by the GPU which leaves math, sound, ai, data management. Math is the only really big item left and if you are running physx most of that is gone anyway, sound is not really relevant and you end up with ai and general data management.

AI has a big benefit due to the amount of actors on the scene but I personally dont see this as a big selling point compared to higher core performance because the benefit of more cores is limited since the more cores you have the bigger data synchronization overhead is needed a 4/8 setup compared to a 8/16 is negligible for games.

Also we should not forget about economics of scale, most game engines will focus on the larger segment so until a very large part of the market runs 8/16 we wont really see any gains at all.

However AMD still has a hit on their hand in the server rooms and this is NOT a small market, just not something I really care that much for though.
 
The French website CPC Hardware has posted what they claim to be AMD Ryzen performance benchmarks. If these numbers are to be believed, the Ryzen's performance looks pretty damn good considering this is an engineering sample in an unknown motherboard running games and benchmarks that can't possibly be optimized for the CPU - motherboard combo.

They ran an A0 engineering sample of Zen and from the New Horizon event there still busy ironing out the kinks on the model they used at that presentation.
AI has a big benefit due to the amount of actors on the scene but I personally dont see this as a big selling point compared to higher core performance because the benefit of more cores is limited since the more cores you have the bigger data synchronization overhead is needed a 4/8 setup compared to a 8/16 is negligible for games.
Also we should not forget about economics of scale, most game engines will focus on the larger segment so until a very large part of the market runs 8/16 we wont really see any gains at all.
However AMD still has a hit on their hand in the server rooms and this is NOT a small market, just not something I really care that much for though.

I would say that the opposite is true , gaming is being held back by the API and the complacent gaming industry where they would put profit over anything else. The development cycle of a game tends to be around 5 years. The way the industry has been operating was always with a keen sense of biggest common denominator approach.

Without this gaming should move forward to see dual cores still operate says enough about the leaps in technology versus the need for better engines were talking about almost a decade now and they wonder why PC market is not selling as well as it used to.
 
Back
Top