AMD RYZEN Demo Event - Beats $1,100 8-Core i7-6900K, With Lower TDP

[rruff]

For example, you will find plenty of people saying that a i3 is superior to an 8 core FX. Nothing more.

That would be me! An i3 performs better (faster) in any task that uses 4 threads or less (which is the great majority) and consumes a lot less power ($) doing it.

So tell me I'm a fanboy. I'd happily get AMD if it made sense. I *was* an AMD fan until it stopped making sense ~8 years ago. I really hope it will make sense again, because we will all be better off. Competition is good.

 

[rruff]

If a RYZEN 8 core 16 thread monster came in at about ~$300 to ~$400 USD retail I would be very interested and tempted. However, if Intel were to drop prices on the Core i7 6900K to the same price range it would be easier and simpler for me to just buy the Core i7 6900K because its a drop-in upgrade. That is a big "if" though.

Expect Intel to match or beat AMD on $/performance and to introduce chips that are superior to anything AMD makes where they will get higher margins.

I could very well be wrong, but I doubt Intel has been really been giving us their "best" the last several years. They simply haven't needed to.

 

[dalekdukesboy]

Just want this to hurry up already hit the shelves and blow up the market. I'm gagging on the high end chip prices for x79 nevermind x99! I think it's almost an understatement to say Intel hasn't been giving it it's best the last several years...why it has no need. Just give small incremental improvements while selling for crazy money because AMD just isn't even in the picture. So yeah, I'm kinda anxious for that crap to end and competition and innovation to be brought back into the market again.

 

[Shihab]

For those who are wondering about IPC judging from the multi threaded tests, the only way possible that Ryzen's IPC is much worse than Broadwell E, is if SMT is much better than HyperThreading ... 8c16t broadwell e has tons of cache and HT is rather mature/refined after all these years, so the IPC should be competitive.
About OC potential, 3.4 GHz base clock is nothing to sneeze at for a 8c16t part ... and with total power consumption lowered, the only real question that remains: Will the chip be as resistant (to electromigration) as intel's at temps over 80C?

Zen's memory hierarchy is slightly different from BW-E's, the latter's cache is only larger on the upper level. I've always had the illusion that L3 cache's size is practically irrelevant for common use (Raytracing and encoding included), can't find a confirmation for it yet though. But it remains that cache size -in general- isn't necessarily a win for Intel or that it helped cover up lower IPC compared to Zen.
Other than that, I agree. It isn't plausible that AMD's SMT application would reach intel's efficiency. Zen's IPC might prove better.

Nitpicking: Hyperthreading is an SMT application, the term is non-brand specific general jargon.

 

[BiggieShady]

But it remains that cache size -in general- isn't necessarily a win for Intel or that it helped cover up lower IPC compared to Zen.

I never meant to compare cache sizes, I meant that the L2+L3 cache is so abundant in Broadwell E so blender/handbrake multithreaded tasks scale almost perfectly ... which means we can somewhat draw conclusions about single thread performance of zen ... it can't be worse unless AMD SMT is more efficient than HT

Nitpicking: Hyperthreading is an SMT application, the term is non-brand specific general jargon.

Yeah, it truly is general term, when I said SMT I meant AMD SMT

 

[P4-630]

AMD RYZEN ZEN 8 Core / 16 Thread CPU Cinebench R15 and Fritz Chess Benchmarks Leaked – Compared Against i7-7700K and i7-6900K

http://wccftech.com/amd-ryzen-zen-cpu-benchmarks-leak/

 

[Eknex]

AMD RYZEN ZEN 8 Core / 16 Thread CPU Cinebench R15 and Fritz Chess Benchmarks Leaked – Compared Against i7-7700K and i7-6900K

http://wccftech.com/amd-ryzen-zen-cpu-benchmarks-leak/

 

[Vayra86]

WCCFTech should just be ignored imo. The amount of BS they spread around the internet is borderline criminal.

 

[GoldenX]

Is there any rumor on lower/lowest end parts, and the TDP support for the different chipsets?
A quad core (Ry)Zen on the lowest chipset (as I understand, that means no chipset at all, just the Zen SoC) should be extremly cheap.

 

[BiggieShady]

Gotta have one

 

[dalekdukesboy]

I thought that was a picture of all their current chips...just add water to perk up a few hundred mhz and you can still only dream of touching Intel's head Watermelon's performance

 

[anubis44]

But what if in 1 year, what AMD is offering is faster than a 6950X for the same price, or the same performance for like $400?

He's still planning to buy Intel anyway, even if it's slower and more expensive.

 

[dalekdukesboy]

He's still planning to buy Intel anyway, even if it's slower and more expensive.

Kinda tough to criticize someone who said something plausible to reply to that after you dig up a quote from 5 pages and days ago...You're sadly defining troll and faboi with one two sentence post, doubt that's your intent but that's your end result.

 

[R0H1T]

Not sure if this has been posted anywhere else, anyway

This is RYZEN (engineering sample)

 

[Caring1]

So basically it is better at calculations and crunching, but still beaten soundly in other areas.

 

[R0H1T]

So basically it is better at calculations and crunching, but still beaten soundly in other areas.

Say what

The tests were done with a clock speed disadvantage of around 5~10% for ryzen, as compared to 6900K, not to mention quad channel memory for BDW-E. So far what I gather, from this review, is that ryzen is between 5960x & 6900k in terms of performance & that with the final production silicon still a few weeks (months?) away. If the auto OC feature works as advertised & ryzen is not an OCing dud, it should sell out quickly provided the price isn't exorbitant.

 

[P4-630]

Note translated with google translate:

https://translate.google.com/transl...lijke-amd-ryzen-benchmarks&edit-text=&act=url

Original article:
https://nl.hardware.info/nieuws/50470/eerste-onafhankelijke-amd-ryzen-benchmarks

 

[bug]

Say what

The tests were done with a clock speed disadvantage of around 5~10% for ryzen, as compared to 6900K, not to mention quad channel memory for BDW-E. So far what I gather, from this review, is that ryzen is between 5960x & 6900k in terms of performance & that with the final production silicon still a few weeks (months?) away. If the auto OC feature works as advertised & ryzen is not an OCing dud, it should sell out quickly provided the price isn't exorbitant.

I'm sure the final silicon will be twice as fast and need half the electrical power

 

[R0H1T]

I'm sure the final silicon will be twice as fast and need half the electrical power

Sure, I'll take your word for it

 

[Vayra86]

What I find far more fascinating is that you can see a 68% perf increase at 75% of the clocks of an FX 8370 while keeping a very low power draw. That's massive progress.

If AMD can do this with a completely fresh architecture on an ES and place that this close to Intel's top end CPU that's been seeing no less than 3-4 refreshes, imagine the refresh a year from now.

 

[EarthDog]

What I find far more fascinating is that you can see a 68% perf increase at 75% of the clocks of an FX 8370 while keeping a very low power draw. That's massive progress.

If AMD can do this with a completely fresh architecture on an ES and place that this close to Intel's top end CPU that's been seeing no less than 3-4 refreshes, imagine the refresh a year from now.

What.... a piledrver to bulldozer increase? (None)

Honestly, I have zero expectations of an amd refresh past this so.... who the hell knows what they put out after this. We don't even know what 'this' is, nonetheless it's next iteration..

 

[Vayra86]

What.... a piledrver to bulldozer increase? (None)

Honestly, I have zero expectations of an amd refresh past this so.... who the hell knows what they put out after this. We don't even know what 'this' is, nonetheless it's next iteration..

Let's not keep putting salt in that old wound It's painful to think about that, hopefully AMD noticed that too and they won't repeat it ^^

Gaming benchmarks: 1% below an i5 that runs at 3.5/3.9. If Zen can push similar clocks, it will be faster than the i5 in IPC. If that is consistent, Im officially impressed.

 

[EarthDog]

Keep wishing...

Point is, we don't know know what this is. Nonetheless the next zen...lol.

 

[P4-630]

After Ryzen will be available I think it takes them many years for a good successor again.

 

[Vayra86]

After Ryzen will be available I think it takes them many years for a good successor again.

Likely, but it will be peanuts to get better yields and better clocking CPUs. It's very normal and is essentially what Intel's been doing ever since they released something on 14nm. Small IPC gains, small clock bumps, nothing extraordinary.