Leaked AMD Ryzen Benchmarks?

CSI_PC said:
If XFR can manage 10%, would you feel that is still bad?
Click to expand...
If XFR's impact was 10% then Ryzen ends up only looking as good on IPC as Ivy/Sandy [assuming leaks are correct, of course]. And while for a first arch in the series it is pretty damn good, it would make me personally think twice whether to get Ryzen rig or not in the summer, since i am cooling limited in the first place. If it is just 100mhz (that's the leeway Passmark leaks have in regards to possible clocks between turboless ES and turbo enabled QS), then it makes XFR overhyped, but Ryzen as a whole stays pretty appealing. And if you meant to mean that XFR enabled on retail only, then it would be alright of course, as long as it does not overvolt the hell out of CPU, like AVFS on Bristol Ridge sometimes does.
 
CSI_PC said:
If XFR can manage 10%, would you feel that is still bad?
Maybe splits opinion but on this node to me that would be pretty reasonable, although we need voltage-frequency-core numbers and behaviour to make a more informed decision, so I appreciate your response is only based upon the basics.
Cheers
Click to expand...

If XFR can take a stock 1800X and boost to 4.4ghz, I'd call that a good feature for people not wanting to mess around too much with the nitty gritty while not making the XFR price increase a complete waste of money for people who don't like playing in BIOS.
 
Magic Hate Ball said:
If XFR can take a stock 1800X and boost to 4.4ghz, I'd call that a good feature for people not wanting to mess around too much with the nitty gritty while not making the XFR price increase a complete waste of money for people who don't like playing in BIOS.
Click to expand...
I tend to think the 1800X is going to be a bit shy of 4.4GHz all cores on air, I just cannot get out of my head it will need similar voltages to Broadwell-E 8C/16t part and above 1.35V.
Cheers
 
lolfail9001 said:
If XFR's impact was 10% then Ryzen ends up only looking as good on IPC as Ivy/Sandy [assuming leaks are correct, of course]. And while for a first arch in the series it is pretty damn good, it would make me personally think twice whether to get Ryzen rig or not in the summer, since i am cooling limited in the first place. If it is just 100mhz (that's the leeway Passmark leaks have in regards to possible clocks between turboless ES and turbo enabled QS), then it makes XFR overhyped, but Ryzen as a whole stays pretty appealing. And if you meant to mean that XFR enabled on retail only, then it would be alright of course, as long as it does not overvolt the hell out of CPU, like AVFS on Bristol Ridge sometimes does.
Click to expand...

Yeah reviews definitely will need to get a grip on the XFR/cores/volt/clock behaviour and comparison between 6C/12T and 8C/16T, also for both normal boost operation and when manually OC'd.
Still if the 1800X can sustain 4.2GHz all cores on air I am interested if the 6C/12T Cinebench multi-thread result ends up being an indicator of overall performance relative to Intel HEDT.
Cheers
 
I don't believe any of these benches yet....

But... I might jump in at the high level after a few months if the threaded benches are nice.

You guys going on about single thread performance are so 1990s.
 
Spirit_Retro said:
I don't believe any of these benches yet....

But... I might jump in at the high level after a few months if the threaded benches are nice.

You guys going on about single thread performance are so 1990s.
Click to expand...


Yep, once you know the single threaded performance and the number of cores you can extrapolate most other things.


The single threaded performance is the basis of all CPU performance.
 
AMD is back, let's not get it twisted, AMD is back and matching INTEL and there's no other way to look at it. Sure you can nit pick here and there, but grand scheme of things, put up an INTEL CPU at 3.4GHz and compare it to 1800X and AMD will win. I can prove this, I have seen this and you all will this coming week.

AMD has extremely good memory bandwidth efficiency, better than INTEL's even in dual channel mode. So I wouldn't say the IMC is weaker per se, yes it's over 1GHZ lower than what INTEL can manage on air cooling (4266MHZ DDR4 vs 3200MHz max on AMD), but weaker is to undersell what AMD has managed here.
So in dual channel mode @ 3200MHZ, CL15 you can get up to 50GB/s memory writes in AIDA 64 which is damn near theoretical limit and that's at CL15.
AMD IMC always runs a command rate of 1T/1N. You can't change it at all and you also can't use really use the tight timings that we can right now on INTEL platforms. Those 12-12-12-26 B-die timings are not going to happen, but no matter, you'll still whip most 7700K systems that aren't tuned to death.

Overclocking this platform is crazy complicated and unrewarding. The clock speeds are low and if you have dreams of a 4.5GHz 24/7 setup, best forget it. In fact most of us will run 3.8 to maybe 4GHz on all cores if we're lucky for 24/7. 1.4V at 4GHz = 90'C load temp with air cooler, so be mindful of that. We need new coolers for the most part because the CPU height is not the same as previous CPUs amongst other changes.

For Haswell-E 5960X to match or beat 1800X (3.6~ 3.7GHzGHz) you'll need about 4.5GHz in the multithreaded programs as Ryzen is ridiculously fast at this beating 6950X sometimes.
Click to expand...
 


Well that kind goes along with what I have been saying about frequency and air cooling and multi-threaded performance.
 
razor1 said:
Well that kind goes along with what I have been saying about frequency and air cooling and multi-threaded performance.
Click to expand...

It is a beast but the beast is in a cage for now, at its clocks it can still handsomely run rings aroung my 5960X ST and MT, he says in anything that involves Rendering, this CPU can run against powerful Xeons and hold its own. I knew about 4 weeks ago that AMD had issues pushing this over 4ghz, at 4ghz it will run like the wind but unless you have high end watercooling that is all you will get.
 
Source? Also, the 3rd paragraph sort of implies that it requires crazy voltage at stock frequency. What?
 
lolfail9001 said:
Source? Also, the 3rd paragraph sort of implies that it requires crazy voltage at stock frequency. What?
Click to expand...

The souce is an insider from ASUS, he is a internationally published tech writer for 20 years now and a professional overclocker.
 
OrangeKhrush said:
The souce is an insider from ASUS, he is a internationally published tech writer for 20 years now and a professional overclocker.
Click to expand...

He mentioned that Voltages are high closer to 4+, these chips are not designed for clockspeed, while the process is so raw, he did say that subsequent steppings will improve but this is not a high frequency chip, this is a high performance chip and a grinder.
 
OrangeKhrush said:
The souce is an insider from ASUS, he is a internationally published tech writer for 20 years now and a professional overclocker.
Click to expand...
I'll take your word for it, it's not like we won't find out in a week.
OrangeKhrush said:
He mentioned that Voltages are high closer to 4+, these chips are not designed for clockspeed, while the process is so raw, he did say that subsequent steppings will improve but this is not a high frequency chip, this is a high performance chip and a grinder.
Click to expand...
Well, looks like the initial rumor of Broadwell-E clocks was closer to reality, then. Well, maybe Intel can mount a proper answer with Skylake-X then. I still find the thought of 4Ghz single core turbo chip needing 1.4V for such frequency weird as hell. Though it is generally in line with AMD's habit of pushing clocks to the limit on high end SKUs.
 
lolfail9001 said:
I'll take your word for it, it's not like we won't find out in a week.

Well, looks like the initial rumor of Broadwell-E clocks was closer to reality, then. Well, maybe Intel can mount a proper answer with Skylake-X then.
Click to expand...

I am still fine with this as I kind of knew it would be somewhere between haswell and broadwell on raw throughput, I did expect good SMT but I was pleasently surprised as I undershot the real SMT numbers by a bit. Clock wise I didn't have much hope beyond 4Ghz as max stable. I knew these chips would all be low frequency high performance parts with maybe the 4C SKU's, but it appears that it shares the same limits despite less resources.

Zen is good, its weakness is that it can't match intel in a drag race, it can however in an arm wrestling match. Basically AMD went from K7 to Netburst to K7(obviously not exactly K7 but low frequency parts for high performance).
 
if AMD can push 4ghz all core clock with high end cooling. I say that is not bad for first gen parts. They focused on performance, clocks can come as the time goes on. 4ghz top is not bad, its only 400mhz or so below my 6850k maxing out at 4.4ghz. We all knew intel had process advantage. If lacking a few 100 mhz on top oveclocks is the only thing amd has to worry about I say thats a good thing.
 
Last edited:
NKD said:
if AMD can push 4ghz all core clock with high end cooling. I say that is not bad for first gen parts. They focused on performance, clocks can come as the time goes on. 4ghz top is not bad, its only 400mhz or so below my 6850k maxing out at 4.4ghz. We all knew intel had process advantage. If lacking a few 100 mhz on top oveclocks is the only thing amd has to worry about I say its a good thing to worry about.
Click to expand...

Thats what I said, Performance that is well up and competitive with Intels top uarches is a good showing, I think AMD laid the foundations, now it is just fine tuning.
 
Zarathustra[H] said:
Yep, once you know the single threaded performance and the number of cores you can extrapolate most other things.


The single threaded performance is the basis of all CPU performance.
Click to expand...

An important component, not the basis. Obviously AMD would not be up there in SMT without improved IPC, nonetheless what is quite impressive is the Superiority of AMD SMT implementation.
 
OrangeKhrush said:
He mentioned that Voltages are high closer to 4+, these chips are not designed for clockspeed, while the process is so raw, he did say that subsequent steppings will improve but this is not a high frequency chip, this is a high performance chip and a grinder.
Click to expand...
Can you ask him if that was with the supplied Wraith air cooler or also with the AM4 Noctua D15 cooler, or something comparable to the D15.
Also is this his experience with both the 6C and 8C or specifically one of them.
Thanks
 
I would like to see the percentages that OC over 4ghz and those that maybe won't. He could have not been lucky with the silicon lottery. Over time here we will be able to see what the numbers are unless Kyle has some Asus inside information on that.
 
noko said:
I would like to see the percentages that OC over 4ghz and those that maybe won't. He could have not been lucky with the silicon lottery. Over time here we will be able to see what the numbers are unless Kyle has some Asus inside information on that.
Click to expand...

Kyle and all the other tech review site owners are literally sitting on their hands wishing they could spill their guts and destroy or support all the leaks (whatever the case may be).

The only thing I'm convinced of so far is that the pricing landscape is about to change. Buy nothing.

We just have to wait.
 
Simplyfun said:
Kyle and all the other tech review site owners are literally sitting on their hands wishing they could spill their guts and destroy or support all the leaks (whatever the case may be).

The only thing I'm convinced of so far is that the pricing landscape is about to change. Buy nothing.

We just have to wait.
Click to expand...

You mean the landscape will change radically on February 28th when AMD puts Intel in a vise profit wise.
 
Simplyfun said:
Kyle and all the other tech review site owners are literally sitting on their hands wishing they could spill their guts and destroy or support all the leaks (whatever the case may be).

The only thing I'm convinced of so far is that the pricing landscape is about to change. Buy nothing.

We just have to wait.
Click to expand...

Yeah, no need to go all crazy. Two weeks from now we'll have all the answers.

(Note: I'm just a humble news poster. I have no inside info or special powers other than the ability to post news, this is just speculative based on what we know to date)
 
Simplyfun said:
Kyle and all the other tech review site owners are literally sitting on their hands wishing they could spill their guts and destroy or support all the leaks (whatever the case may be).

The only thing I'm convinced of so far is that the pricing landscape is about to change. Buy nothing.

We just have to wait.
Click to expand...

I think you're right. A certain Editor/Site Owner around here was pretty active in the AMD threads and then it dropped off a few days ago. I bet a NDA is in play.
 
You must be kidding. We are supposed to freeze buying because Intel announces some lame skus that won't be available for months? What hutspa! I am buying day one at my local Micro Center. To hell with Intel and its lap dogs. Chicken Little says "The sky is falling. Chicken little calls out Freeze sky! A few mantras, a few metaphysical mutterings, will do. Meanwhile Humpty Dumpty has fallen off the wall. And all Intels trolls and all its clay feet knights cant put Humpty back together again.
 
Last edited:
os2wiz said:
You must be kidding. We are supposed to freeze buying because Intel announces some lame skus that won't be available for months? What hutspa! I am buying day one at my local Micro Center. To hell with Intel and its lap dogs. Chicken Little says "The sky is falling. Chicken little calls out Freeze sky! A few mantras, a few metaphysical mutterings, will do. Meanwhile Humpty Dumpty has fallen off the wall. And all Intels trolls and all its clay feet knights cant put Humpty back together again.
Click to expand...

Well, that's the job of the marketing department of any company that is behind their competitor and have something in the works. To - as best as they can - convince buyers to postpone their purchases because they have something great on the horizon. Same thing AMD has been doing for several months with Zen now :p

I agree. I'd rather give AMD my money sooner than wait for something from Intel later, but I've never been one to rush out and buy on day one. There is no merit to being "first" to own something. In fact, early adopters are the ones who always have the most problems. If AMD lives up to these latest expectations, I'll buy when I am ready. Maybe after the first stepping revision, and first revision update on the motherboard I want, so I know they have ironed out the kinks :p
 
I'm still running an Ivy Bridge i5 @ 4.4 ghz. It's been a great CPU, but I'm hoping AMD gives me a reason to jump ship and grab 1700x. Give me a reason AMD, I'm ready to build a SFF machine.
 
I am personally looking forward to retiring my very old 8350. Looks like Zen will be what I want. Which is lots of cores, decent clock speed, great multi threaded performance and very good single threaded performance. Might not beat Intel on single threaded performance but it's close enough.
 
jbean7457 said:
I'm still running an Ivy Bridge i5 @ 4.4 ghz. It's been a great CPU, but I'm hoping AMD gives me a reason to jump ship and grab 1700x. Give me a reason AMD, I'm ready to build a SFF machine.
Click to expand...

I've got an i7-4770K I'm running at 4.3ghz (despite low temps it can't stay stable higher), but I got the whole CPU/Mobo/RAM cheap from a friend and I'll probably resell it for higher than I bought it for after building an 8 core rig.

Main reason is newer I/O and the few apps and games that can utilize more cores.

I do it for fun :)
 
Zarathustra[H] said:
Well, that's the job of the marketing department of any company that is behind their competitor and have something in the works. To - as best as they can - convince buyers to postpone their purchases because they have something great on the horizon. Same thing AMD has been doing for several months with Zen now :p

I agree. I'd rather give AMD my money sooner than wait for something from Intel later, but I've never been one to rush out and buy on day one. There is no merit to being "first" to own something. In fact, early adopters are the ones who always have the most problems. If AMD lives up to these latest expectations, I'll buy when I am ready. Maybe after the first stepping revision, and first revision update on the motherboard I want, so I know they have ironed out the kinks :p
Click to expand...

Common sense dictates you are right. Especially for 4 core and 6 core Ryzens , I expect higher clocked iterations when they get dedicated wafer production.Yes you can occasionally get rev 1.0 mothrerboards that have been disasters especially if a chipset has a nasty bug. But that is not often.
 
OrangeKhrush said:
The souce is an insider from ASUS, he is a internationally published tech writer for 20 years now and a professional overclocker.
Click to expand...

The text only exist 2 places when googled. And that's from you on H and AT.

Also 3200Mhz on a 6700K can reach 49GB/sec in AIDA with 3200Mhz, with CL16.

The command rate comment is BS too. Not to mention the rest of your fairy tales.
 
Last edited:
Anyone know of the chances of Ryzen having the instructions required for 4k Netflix decoding that, supposedly, only exist on Kaby Lake?

Currently that is my main deciding factor of what CPU I am putting into my SFF build. If I know Ryzen has it, I can hold off, but if I don't, I might just build it next week (long holiday), since I don't really need it to do anything else.
 
Shintai said:
The text only exist 2 places when googled. And that's from you on H and AT.

Also 3200Mhz on a 6700K can reach 49GB/sec in AIDA with 3200Mhz, with CL16.

The command rate comment is BS too. Not to mention the rest of your fairy tales.
Click to expand...

Well Cinebench did conclude one of your concerns, clock for clock it is basically haswell in single thread IPC, but SMT is much higher than expected. I think it is more your hostile interjections that seem to be the only thing getting old and stale.
 
OrangeKhrush said:
Well Cinebench did conclude one of your concerns, clock for clock it is basically haswell in single thread IPC, but SMT is much higher than expected. I think it is more your hostile interjections that seem to be the only thing getting old and stale.
Click to expand...

When was the CB number verified? And hostile for what, saying your BS is BS? Prove it if you want it otherwise :)
 
chenw said:
Great... thanks.

EDIT: Colour me ignorant, but the instruction for that is part of iGPU?
Click to expand...

He refers to the video decoder in the graphics part. Ryzen is without a graphics part. But if you buy a pascal based card or a GM106 you get full hardware decode of the 4K movies. Anything else besides Kaby Lake is hybrid support at best.
 
Shintai said:
He refers to the video decoder in the graphics part. Ryzen is without a graphics part. But if you buy a pascal based card or a GM106 you get full hardware decode of the 4K movies.
Click to expand...

Actually, it isn't with the movie itself. At least I don't think it is.

Netflix, on their website, explicitly states that 4k streams requires Kaby Lake CPU and either Edge or its Win 10 app (the latter, understandable, the former, might have some MS-Netflix love, since it's the only browser to support 1080p stream, otherwise all other browsers only get 720p). which apparently has something to do with a particular instruction about encoding/decoding its DRM (I left this bit out, my apologies).

I have no clue whether the required instruction is CPU or iGPU, but I assume CPU because Netflix doesn't state any other possible alternatives (even dGPU side).

It doesn't seem to have anything to do with decoding the media file itself.

EDIT: Source: https://help.netflix.com/en/node/55763

The requirements are ripe for conspiracy theories too.
 
Last edited:
You must log in or register to reply here.
Back
Top