Leaked AMD Ryzen Benchmarks?

Shoot I sure hope this isn't a flop guys. I just scored the amazon 5% cash back visa for prime members and with it a 70$ gift card. Applied the latter to my pre-order. Gonna be some bad feels over here much sadness if I read actual legit full reviews and the 6850K ends up looking better.
 
So with turbo activated that would only be a10% advzntage. If you deduct 10% off Kaby Lake IPC you get A drop higher than Broadwell for Ryzen IPC.
Not sure you are going to get a satisfying reponse as his post was further speculations used as evidence, quoting someone else saying it must be base clocks....
The one aspect he did not touch upon is whether AMD also ensured multicore support enabled with auto for the 6900K.
However I went further out of my way and checked, that Asus motherboard defaults to multicore enhancement auto meaning it ignores the Intel spec for multicore frequency behaviour and attempt all cores at 3.7GHz and will only reduce this due to throttling, this is not really overclocking (semantics could be argued I guess but more can be said about Intel's Boost Max 3.0 that pushes a core to 4GHz ) because it will never go over the frequency boost limit of 3.7GHz.

So either AMD has substantially beaten Intel's SMT and that by a substantial amount taking base clock vs an all-core 6900K enabled to 3.7GHz (as I mentioned in the past this is probably how it was working and throttling due to cooler) or much more likely and in realms of possibility it was not fixed at base clocks but still beats the 6900K in terms of SMT by a good figure.
Also worth noting the results are sort of in-line allowing for variance with the good review sites, such as Tom's who stated they saw all cores running at 3.7GHz for their 6900K
Anandtech had a better multithread score but worth noting they are using a decent cooler, where AMD event used the default coolers for these CPUs (so both are potentially throttling just by how much is the question although the difference is not that large and under 5% between AMD event and Anandtech).

I appreciate you understand a lot of this but just wanted to expand on his context and attempts to skew using non-verified information.
Cheers
 
Last edited:
Not likely in my opinion. Don't let anxiety rule you.



Shoot I sure hope this isn't a flop guys. I just scored the amazon 5% cash back visa for prime members and with it a 70$ gift card. Applied the latter to my pre-order. Gonna be some bad feels over here much sadness if I read actual legit full reviews and the 6850K ends up looking better.
 
People are mostly concerned with CPU performance, not a small power difference. Actual power draw can vastly differ from the rated TDP as well.
We don't know how AMD is choosing to rate TDP on this CPU either.

It could be like how they chose to do it with excavator or they could be doing it like Intel.
 
Not sure you are going to get a satisfying reponse as his post was further speculations used as evidence, quoting someone else saying it must be base clocks....
The one aspect he did not touch upon is whether AMD also ensured multicore support enabled with auto for the 6900K.
However I went further out of my way and checked, that Asus motherboard defaults to multicore enhancement auto meaning it ignores the Intel spec for multicore frequency behaviour and attempt all cores at 3.7GHz and will only reduce this due to throttling, this is not really overclocking (semantics could be argued I guess but more can be said about Intel's Boost Max 3.0 that pushes a core to 4GHz ) because it will never go over the frequency boost limit of 3.7GHz.

So either AMD has substantially beaten Intel's SMT and that by a substantial amount taking base clock vs an all-core 6900K enabled to 3.7GHz (as I mentioned in the past this is probably how it was working and throttling due to cooler) or much more likely and in realms of possibility it was not fixed at base clocks but still beats the 6900K in terms of SMT by a good figure.
Also worth noting the results are sort of in-line allowing for variance with the good review sites, such as Tom's who stated they saw all cores running at 3.7GHz for their 6900K
Anandtech had a better multithread score but worth noting they are using a decent cooler, where AMD event used the default coolers for these CPUs (so both are potentially throttling just by how much is the question although the difference is not that large and under 5% between AMD event and Anandtech).

I appreciate you understand a lot of this but just wanted to expand on his context and attempts to skew using non-verified information.
Cheers

I think you are asking for to much, skewing, getting it wrong by a lot and never admitting to as much is the order of the day. You are going to be like a dog chasing tail, round and round and round with no joy. Most people on this side of the forum have in some way embraced the possibility that Ryzen is what Sandybridge was to Intel and the PC market in general, it is nothing short of brilliant engineering. But like everything it will be met with its fair share of detractors. One such in our local domain complained about HPC, he is a HPC user and knew from the outset that AMD made sacrifices in AVX throughput because to a gamer or consumer level content creator AVX is not useful at all. There is always going to be knitpicking for faults. It has been 2 years of it and frankly I am jaded of it.

Ryzen = Sandy, just to save yourself a lot of misery.
 
Someone else pointed out that there isn't a Ryzen etching on the cpu. I'm under the impression that the press chips and retail all have it etched.


can't really tell with the way the lighting is , the CPU etching could just be washed out due to the lighting nothing else.
 
can't really tell with the way the lighting is , the CPU etching could just be washed out due to the lighting nothing else.

I just went through the vid, hmm he states he got the cpu but not the motherboard and he had to wait? Hmm...the press kits include am4 mobos. I guess AMD could have forgotten to include it....
 
Someone else pointed out that there isn't a Ryzen etching on the cpu. I'm under the impression that the press chips and retail all have it etched.

I can't quite make it out in the video, but it appears from what I can see to start with "ZD" but the etching looks identical to the QS samples we've seen floating around.
 
The guy that made the video, (company for that matter) is a system builder so, if he got a retail sample, and not review sample, they wouldn't have the motherboard. Don't see why they would lie if they won't be able to sell the product...... What is the use for them start promoting sales of their own systems without having the hardware.... Kinda defeats the purpose right?
 
I still don't get why Ryzen 8 core chips are being compared to 22 core Xeons.

Wait for the already confirmed Naples chips which go all the way to 32.

Compare on performance, cost, and core count. Comparing an 8 core $500 chip to a 22 core $4500+ chip does none of those things.
 
I still don't get why Ryzen 8 core chips are being compared to 22 core Xeons.

Wait for the already confirmed Naples chips which go all the way to 32.

Compare on performance, cost, and core count. Comparing an 8 core $500 chip to a 22 core $4500+ chip does none of those things.

Who compares it to them?
everyone compares it against 6900K\6800K and 7700K.
 
OK to put it to rest I went back through the Event presentation.
In the footnotes it only mentions locked clocks for the IPC uplift where it was at 3.4GHz, in the footnotes it also has information about the Cinebench test for 1800X and 6900K and no mention of locked clocks.
Furthermore, in the system config last page, it clearly shows for 1800X Base 3.6GHz Boost 4GHz, with the 6900K base 3.2GHz Boost 3.7GHz and boost Tech Max 3.0 active with 4GHz.
Unfortunately it is difficult to deduce much from the power demand figure due to nearly everyone else measuring this with more power demanding applications than Cinebench, and also depends how sensitive-time window-accurate the mains monitor is as well.
The power consumption will not be directly comparable or as a direct correlation to AMDs power demand figures but Anandtech figures here shows how the 6900K behaves in terms of power demand and core usage, they use Prime95 (more of a stress than Cinebench) and showed it was moderately linear (after 2 cores) as core count usage increments; 41W 2 core, 75W 4 cores, 149W All 8 cores.
So at most one can say is that neither were held back in the Cinebench test apart from possibly (but we cannot say either way until reviews) a bit of throttling due to the default coolers on the 1800X and 6900K.
Cheers
 
Last edited:
If AMD is quoting 52% IPC, that is a best case scenario, so why is it hard to figure out best case numbers? Just remember, regardless of who AMD had working on these chips, Intel has an R&D budget that is about the same as AMDs total market cap. That's only recently as well, it was much larger before.

I hope it's at least competitive at price / performance
 
If AMD is quoting 52% IPC, that is a best case scenario, so why is it hard to figure out best case numbers? Just remember, regardless of who AMD had working on these chips, Intel has an R&D budget that is about the same as AMDs total market cap. That's only recently as well, it was much larger before.

I hope it's at least competitive at price / performance
Also the US spends more money per student by far then any other country in the World and the results suck! R&D budget may not indicate success in the end - it may help the chances of success but not necessarily win the day. Intel's performance per generation increase, last generations have not been that great.

Just several more days for the real hard facts to come out.
 
If AMD is quoting 52% IPC, that is a best case scenario, so why is it hard to figure out best case numbers?
Actually, it is not the best case scenario. It is actually one of the worse case scenarios, when they overshoot their initial estimations and are still lagging behind Intel.
Also the US spends more money per student by far then any other country in the World and the results suck
Because you guys have students going into colleges with degree in religious experiences instead of knowledge. Granted, that neatly translates to Intel with it's spending all over the place.
 
Actually, it is not the best case scenario. It is actually one of the worse case scenarios, when they overshoot their initial estimations and are still lagging behind Intel.

Because you guys have students going into colleges with degree in religious experiences instead of knowledge. Granted, that neatly translates to Intel with it's spending all over the place.
You forgot to mention safe zones, gender neutral (pick what sex you think you are) rest stops for those tender ears and bottoms. Also the right to protest, destroy and burn up stuff as well and use crayons for therapy if the president that was elected is so shocking to you.
 
Actually, it is not the best case scenario. It is actually one of the worse case scenarios, when they overshoot their initial estimations and are still lagging behind Intel.

Because you guys have students going into colleges with degree in religious experiences instead of knowledge. Granted, that neatly translates to Intel with it's spending all over the place.
52% over Pilderiver, 64% improvement ofer Excavator.
 
52% over Pilderiver, 64% improvement ofer Excavator.
Yeah, the irony if this does not fail to escape me.
The point being SPEC is one of benches AMD still loses, though i do not know if AMD included libquantum in their presentation.
 
i heard that the overclocked 1800x can beat the 7700k in single thread. I'm assuming both cpu's are running at same clocks
 
i heard that the overclocked 1800x can beat the 7700k in single thread. I'm assuming both cpu's are running at same clocks
Clock to clock comparison for single thread, 2 threads etc. would be most enlightening. Hope someone does that.
 
so if the 91 watts is a little to liberal on intel does that mean that the 65 watts its too conservative on AMD?

The Toms one is OC to 4.5Ghz on all cores if you didn't notice. A 6700K/770K in anything but 256bit AVX/FMA benches tends to be below 65W. Also ask yourself, why is AMDs 65W chip shipping with a 95W cooler, when they do have 65W coolers for other 65W chips? :D
 
Not sure you are going to get a satisfying reponse as his post was further speculations used as evidence, quoting someone else saying it must be base clocks....

He was quoted in a reply explaining him that he was wrong and that turbo was activated. It is not speculation. It is obvious that turbo was activated, otherwise the scores reported by WCCFTECH don't make sense.
 
Yup that's why competition is so important.

That has been discredited multiple times and in multiple forums. I will copy and paste what a mod wrote in another forum:

that Intel 'price cut' isn't really.

Its a Micro Center Sale

you know, Micro Center, right? the brick and mortar store that ALWAYS discounts CPUs and select mobos to get you in the door because everything else is more expensive than Newegg?

That's what that rumor on deep discounts is. If you're near a Micro Center, its great. But it's nothing new from them and it is definitely not an Intel conspiracy!
 
The Toms one is OC to 4.5Ghz on all cores if you didn't notice. A 6700K/770K in anything but 256bit AVX/FMA benches tends to be below 65W. Also ask yourself, why is AMDs 65W chip shipping with a 95W cooler, when they do have 65W coolers for other 65W chips? :D

AMD's 65w chip ships with a 65W cooler.
 
If AMD is quoting 52% IPC, that is a best case scenario, so why is it hard to figure out best case numbers? Just remember, regardless of who AMD had working on these chips, Intel has an R&D budget that is about the same as AMDs total market cap. That's only recently as well, it was much larger before.

I hope it's at least competitive at price / performance

It is not a best case scenario, but a kind of average. For CB15, AMD reports 58% higher IPC than Excavator. For SPEcint, AMD reports a 64% higher IPC than Excavator. Myself used the 64% figure to estimate SPECint scores for Zen here

https://hardforum.com/threads/leaked-amd-ryzen-benchmarks.1920876/page-27#post-1042840661
 
Now it is stock cooling fault. OEM's just ship stock, not like OEM's sell 6900K's but nevertheless any 6700 or 7700 they do ship is stock cooled, if that is not sufficient it is not the problem of AMD for using stock fans. 60BN conglomerate cannot ship better cooler, to sad to hear about your troubles.
 
Now it is stock cooling fault.
I'll be honest, it does look suspect, especially since Zeppelin should really have no issues with heat transfer with nearly 200mm^2 die, solder and supposedly 65W TDP.
1700X beating 7700K in single-thread.
Didn't we establish it before that CPU-Z bench is a bit flawed? Also, where is the validation screen?
 
I'll be honest, it does look suspect, especially since Zeppelin should really have no issues with heat transfer with nearly 200mm^2 die, solder and supposedly 65W TDP.

Didn't we establish it before that CPU-Z bench is a bit flawed? Also, where is the validation screen?

So what you suggest is that AMD handicaps themselves by allowing their CPU's to possibly throttle (maybe they just don't hit throttle limits) but give Intel the best possible environment. It is an out the box vs out the box test and if AMD's CPU and cooler is better, well that is tough cookie.
 
So what you suggest is that AMD handicaps themselves by allowing their CPU's to possibly throttle (maybe they just don't hit throttle limits) but give Intel the best possible environment.
I have no issue with AMD using a better cooler, my concern lies in the area of them pulling a Polaris trickery, because if i am to put Ryzen in my build, i do not want my itx box to turn into overheating crapshoot with just locking cores to their single core turbo.
 
I have no issue with AMD using a better cooler, my concern lies in the area of them pulling a Polaris trickery, because if i am to put Ryzen in my build, i do not want my itx box to turn into overheating crapshoot with just locking cores to their single core turbo.

From one ITX user to another, even a i5 4460 cannot survive a stock cooler in an ITX setup, the Graphics Card heat dissipation is to close to the CPU and it pumps all that hot air straight into the CPU fan. If you are running ITX you really should be on AIO or solid HS/F. I don't see how a GPU can in anyway be related to how Ryzen operates, it was hitting 4Ghz on stock, Polaris was very different and GPU's in general are far mor sensitive to throttling or putting themselves into imposed V-Sync. If a CPU hits thermal throttle in a bench most times you black screen. Intel even endorse their chips to be able to go over TDP. I don't think that the 6900K was handicapped at all.
 
Back
Top