TR Forums

These slides are not 100% vetted but given that I'm writing this less than 24 hours before the big "New Horizons" event, they might be legit:
http://videocardz.com/64741/ryzen

Apparently Zen's non-codename product name is... RyZen... not sure I particularly like it but I'm 100% pronouncing it to rhyme with Rye-Bread.

Frankly there's one piece of actually new information in these slides: Apparently the highest end chip is only going to be hitting 3.4GHz at launch with the "+" being in there to denote the fact that you can overclock some of the SKUs.

lol wow. if ryzen is real, that AMD marketing team needs to be fired. now. yesterday.

Getting my Flame retardant pantz ready .. should be a interesting read and demo followed by all sorts if conjecture and bickering.  Popcorn at the ready!

maxxcool wrote:

Getting my Flame retardant pantz ready .. should be a interesting read and demo followed by all sorts if conjecture and bickering.  Popcorn at the ready!

Oh yeah, tomorrow should be... interesting.
If AMD just does demos of these "RyZEN" parts on their own merits and can actually show the maturity to make this a pro-AMD event instead of trying to drag down their competition then I'm willing to give them the benefit of the doubt.
If, however, this turns into some whacko "male bovine excrement" fair where there are totally unrealistic and contrived comparisons made to fool people into thinking that RyZEN is going to destroy a 10 core Intel processor, then they'll have another thing coming.

DancinJack wrote:

lol wow. if ryzen is real, that AMD marketing team needs to be fired. now. yesterday.

AMD has a marketing team?

The highway end chip (max # cores) isn't going to be a clockspeed champion. If their 4-core part Todd out at 3.4ghz, then it will be another chip sold in the bargain bin.

Kretschmer wrote:

The highway end chip (max # cores) isn't going to be a clockspeed champion. If their 4-core part Todd out at 3.4ghz, then it will be another chip sold in the bargain bin.

Posting from your phone, eh?

chuckula wrote:

Apparently Zen's non-codename product name is... RyZen... not sure I particularly like it but I'm 100% pronouncing it to rhyme with Rye-Bread.

It's either "rye-zen" or "risen". The latter is sure to get them into trouble, but maybe at this point they just don't care anymore. I'd like to think it's a joke about their current do-or-die situation in the CPU market, but I doubt that it is.

Pastrami?? Did someone say pastrami??

I love pastrami. That is all.

Captain Ned wrote:

Pastrami??  Did someone say pastrami??

I love pastrami.  That is all.

Finally something that can be agreed upon in the forums!

Redocbew wrote:

It's either "rye-zen" or "risen". The latter is sure to get them into trouble, but maybe at this point they just don't care anymore. I'd like to think it's a joke about their current do-or-die situation in the CPU market, but I doubt that it is.

It's surely the latter. I mean come on guys, how does "rye-zen" make ANY sense?

RyZen isn't too bad a name, but has no memorable aspects as you'd expect from a life and death product.
The first day I reported to AMD, as I entered the building with a VP, he smiled and whispered in my ear, "Prepare to meet the most privileged and incompetent people on the planet".
It's amazing how they all looked and thought just alike.

chuckula wrote:

Captain Ned wrote:

Pastrami??  Did someone say pastrami??

I love pastrami.  That is all.

Finally something that can be agreed upon in the forums!

Only if the meat is hot.

DancinJack wrote:

Redocbew wrote:

It's either "rye-zen" or "risen". The latter is sure to get them into trouble, but maybe at this point they just don't care anymore.  I'd like to think it's a joke about their current do-or-die situation in the CPU market, but I doubt that it is.

It's surely the latter.  I mean come on guys, how does "rye-zen" make ANY sense?

I REALLY hope for AMD's sake it's Rye-Bread (I mean Rye-Zen).
The pretty offensive religious connotations of pronouncing it "risen" is offputting enough.  Not to mention the jokes if it fails to "rise" properly.... yeah, there could be some good material there that will get used extensively if AMD screws the pooch on this one.

just brew it! wrote:

DancinJack wrote:

lol wow. if ryzen is real, that AMD marketing team needs to be fired. now. yesterday.

AMD has a marketing team?

Yap and if i recall it correctly Scott is a part of it.

chuckula wrote:

I REALLY hope for AMD's sake it's Rye-Bread (I mean Rye-Zen).
The pretty offensive religious connotations of pronouncing it "risen" is offputting enough.  Not to mention the jokes if it fails to "rise" properly.... yeah, there could be some good material there that will get used extensively if AMD screws the pooch on this one.

I really don't get why you guys are going to Jesus for just the word risen. I guess I understand why, but it's not where my brain goes. The first definition for rise (risen is past participle i think) is "move from a lower position to a higher one; come or go up." I think it's exactly something AMD would do.

Corned Beef > Pastrami, although I do like both.  

So Ryzen SR7/5/3 will be the full SKU names?  Hopefully it performs on par with Haswell, that would be very good as Skylake-E isn't here yet so they'd be undercutting the octocore X99 chips by a good amount if the rumored 350-500$ prices are anything to go by.

The fact that Vega is outperforming GTX 1080 by 10% in cherry picked benchmarks paints a pretty good picture, it should at least be GTX 1070 caliber.

Am I the only one who's first though was ricin?

I'm surprised I didn't considering I watch Breaking Bad quite a bit.

Some real tests showing Zen (or ryzen) go head to head even with that old reliable, Sandy Bridge, should be enough to make me happy - I'm frugal, that way. Convoluted testing under strangely concocted situations against 1000$ intel CPUs paired with bizzaro-world iGPU or low-midrange GPU combinations that no one buys as such will make me extremely skeptical, if not outright pessimistic that they've screwed up.

We don't give two **** about how it compares to intel extreme CPUs. Give us real tests against REAL intel cpus that people actually buy..

Smells fake to me. Remember the fake slides before the last Zen presentation?

Fakes McFakesterston

Hey, guys, we should all listen to RyZen (reason).

DancinJack wrote:

The first definition for rise (risen is past participle i think) is "move from a lower position to a higher one; come or go up." I think it's exactly something AMD would do.

Well, AMD's CPU business is pretty much dead.

DancinJack wrote:

lol wow. if ryzen is real, that AMD marketing team needs to be fired. now. yesterday.

AMD's marketing should have been fired years ago. All they do is overhyhpe and leave people disappointed when their products actually launch.

chuckula wrote:

Captain Ned wrote:

Pastrami??  Did someone say pastrami??

I love pastrami.  That is all.

Finally something that can be agreed upon in the forums!

I'll eat to that!
Also current rumor puts the 8 core chips and 3.2/3.5 and 3.0/3.2. From a mostly gaming perspective I'm curious how the 6 and 4 core parts will clock. *crosses fingers for 4GHz*

MileageMayVary wrote:

Also current rumor puts the 8 core chips and 3.2/3.5 and 3.0/3.2. From a mostly gaming perspective I'm curious how the 6 and 4 core parts will clock. *crosses fingers for 4GHz*

OK, so where would that put us performance-wise relative to the FX-8350, if we take the "40% better IPC than Excavator" claim at face value?

I don't know what the effective IPC improvement was between Piledriver and Excavator... rough guess of 15%? If we say Excavator had 15% better IPC than Piledriver (and Zen is 40% better than Excavator), this puts Zen at roughly 60% IPC improvement over Piledriver. But if the above rumor is true, top-of-the-line Zen at launch has a stock clock only 80% of FX-8350's. So after all this hand-waving, we supposedly end up with something that performs roughly on par with a 5 GHz octo-core Piledriver part, with a 95W (?) TDP. Not terrible, but not exactly earth-shattering either.

I guess we will just need to wait and see the benchmark results for real applications. Is Zen destined to be a dream or a dud?

just brew it! wrote:

MileageMayVary wrote:

Also current rumor puts the 8 core chips and 3.2/3.5 and 3.0/3.2. From a mostly gaming perspective I'm curious how the 6 and 4 core parts will clock. *crosses fingers for 4GHz*

OK, so where would that put us performance-wise relative to the FX-8350, if we take the "40% better IPC than Excavator" claim at face value?

I don't know what the effective IPC improvement was between Piledriver and Excavator... rough guess of 15%? If we say Excavator had 15% better IPC than Piledriver (and Zen is 40% better than Excavator), this puts Zen at roughly 60% IPC improvement over Piledriver. But if the above rumor is true, top-of-the-line Zen at launch has a stock clock only 80% of FX-8350's. So after all this hand-waving, we supposedly end up with something that performs roughly on par with a 5 GHz octo-core Piledriver part, with a 95W (?) TDP. Not terrible, but not exactly earth-shattering either.

I guess we will just need to wait and see the benchmark results for real applications. Is Zen destined to be a dream or a dud?

I surely hope that for the first time in many years, AMD has been underselling what they've built. 40% on average would barely get them to Sandy Bridge performance, let alone Haswell like many are hoping. I'd take a 4 GHz Sandy Bridge over a 5 GHz Piledriver any day.

chuckula wrote:

The pretty offensive religious connotations of pronouncing it "risen" is offputting enough.  Not to mention the jokes if it fails to "rise" properly.... yeah, there could be some good material there that will get used extensively if AMD screws the pooch on this one.

You're leaking a little bit there. Maybe get that checked?

I'm going to go on record right now and predict that Zen or RyZEN or SR7 or whatever the hell AMD calls these chips will in fact be competitive clock for clock with Broadwell-E, which would be a huge improvement over current FX chips.

I think the combination of SMT on every chip, the added bandwidth of DDR4, the new 20mb of combined L2/L3 cache, and updated chipsets will make Zen a very attractive option. If AMD really releases an 8C/16T part at 3.4ghz that's 200mhz higher clocked than the 6000K.

More importantly is the price, if in fact the 8C/16T part are in the $500 range then that bodes well for the Opteron parts, which should allow one to build a nice 16C/32T workstation relatively cheap.

I think AMD will prove to have a winner on it's hands and this is coming from a guy that built more Intel based systems than AMD based ones and uses Intel.

The 40% is over Carrizo, not AM3 procs.

AMDs newest arch pre-Zen doesn't even have an AM3 version.