Official AMD Ryzen Benchmarks, Reviews, Prices, and Discussion

[french toast]

AMD sells the FX-8350 for $150. Who loses most if AMD has to sell the 1700X at $150?

Intel, but that wont happen anytime soon.
All costs accounted for, including R&D, marketing, ect probably summit ridge cost 100$ to make.

 

[lolfail9001]

INTEL has had a settlement with AMD and the national authority (I quite remember a settlement also in Europe) and promised to do not repeat those actions and in return it obtained a lower punishment. But if INTEL repeats the crime, it will be punished twice.

You can't deny it, because it was punished by at lest USA and Europe. I don't remember the fees, but they were reduced because in exchange INTEL settled with AMD with more than 1 billion dollar and the relaxation of the terms on the x86 license, more notabily the obligation to own a FAB.

Last time around the backdeal was with OEMs, not with end users, that's his point.

 

[Doom2pro]

AMD sells the FX-8350 for $150. Who loses most if AMD has to sell the 1700X at $150?

Doubt if AMD has much inventory of FX parts, mostly just e-tailers with old stock sitting around that are going to suffer.

 

[krumme]

I trust AMD won't be selling the FX-8350 for much longer. It should really only be a $70 CPU anyway.

However, if Intel drops prices so low that the 1700X is forced into such a bracket then anti-trust laws will absolutely be broken. I think it's pretty safe to say that Intel will keep the retail prices rather high and will work by creating huge discounts for partners... in exchange for their loyalty...

I think its a fairy nice performing cpu that is strong where it matters and is most critical. In new games and heavy workloads. I would say 80 or even 90 usd is ok for the performance.

The basic problem is just not the price but that its a toaster. 125w tdp. Serious. Its the double of a r7 1700.

 

[dark zero]

After seeing AMD's scores and prices I can say that Intel royally screwed up with their pricing... and their timing. Seems that AMD succeded at last after a ton of years later. Hope that the success is maintaining.

 

[formulav8]

... and their timing

That's why Intel may be looking to possibly start their old shenanigan's.

Doesn't mean AMD dominates it all, but just enough for Intel to take a hit with a so-called "penalty" charge. When you do not have morals, you weigh the penalty, whether short/long term and see if you're better off.

 

[PrivateCeralion]

After seeing AMD's scores and prices I can say that Intel royally screwed up with their pricing... and their timing. Seems that AMD succeded at last after a ton of years later. Hope that the success is maintaining.

Yes, the Kaby Lake timing was horrible. Who need a 5Ghz 4Core Cpu, if you have a 8Core 65W CPU running at the sweetspot? It would be better for Intel, if Kaby Lake hasn`t published and they counter Ryzen in the Summer with making Skaylake X mainstream.

 

[formulav8]

with making Skaylake X mainstream

I hope its true, but Intel is probably the same ole Intel. I don't expect it I guess.

 

[.vodka]

http://ranker.sisoftware.net/show_r...dae7d4edd5e4d6f082bf8fa9cca994a482f1ccfc&l=en

R7 1800X cache performance

Details for Result ID AMD Ryzen 7 1800X Eight-Core Processor (8M 16T 3.59GHz, 8x 512kB L2, 2x 8MB L3);

Local Rank #819

Position Higher than 92.86% ranked results

Points 93

Score 229.54GB/s

Qualification Excellent Performance

L1 Cache Bandwidth 776.86GB/s
L2 Cache Bandwidth 590.10GB/s
L3 Cache Bandwidth 408.71GB/s

Speed 3.59GHz

16GB Corsair CMK16GX4M2B3000C15 DIMM (2.13GHz) PC24000 (16-16-17-39 4-55-18-6)
Name GigaByte Default string (GigaByte AX370-Gaming 5)



Can we say AMD has finally fixed their cache performance after all these years??

(This also means good caches for Vega!)

edit: That's DDR4 3000 for that Gigabyte motherboard. Glad to see those BIOS updates doing their job!

 

[formulav8]

I have wondered if AMD purposely sandbagged in some performance scenarios honestly.

I think their absolute gaming performance may be lower than Intel, but overall they made a real hit.

Edit: Spelling , etc.

 

[formulav8]

Argg, looks like a double post.

 

[cytg111]

I have wondered if AMD purposely sandbagged in some scenarios honestly.

I thibk their absolute gaming performance could b loer, but oveall

You and me both
https://forums.anandtech.com/thread...use-new-thread.2482739/page-159#post-38651401

 

[GroundZero7]

So, like apparently Cinebench is no longer a representative benchmark, this happened between all the big talking at "new horizons" and changed after the 22/2 NDA lift, it is amazing how people can change their tune.


OKAY PEOPLE LETS MOVE THESE GOAL POSTS OVER THERE.

Isn't that the Arma 3 only section?...I have a problem with that.

 

[lolfail9001]

edit: That's DDR4 3000 for that Gigabyte motherboard. Glad to see those BIOS updates doing their job!

I can't be the only one noticing C15 in the name and 16 in the timings, though.

Also
http://ranker.sisoftware.net/show_r...dae7d4e3d0e2d6f082bf8fa9cca994a482f1ccfc&l=en
Looks like Zen's cache size certainly affects bandwidth, in comparison to 6950X.

 

[DrMrLordX]

Last time around the backdeal was with OEMs, not with end users, that's his point.

For large-scale buyers, sales channels are different now than back in 2005 and earlier. The really big ones (Amazon, Google, MS, etc) are buying through ODMs, and in some cases they may be buying directly. Intel has already started customizing chips for their larger clients. Those customizations may even be baked into Xeons sold to smaller parties without anyone being the wiser.

Don't think Intel is unaware of a large firm's interest in AMD processors. It isn't an absolute certainty that an OEM rep will be on the phone with the procurement department, grooming the client for another sale of Intel-based hardware. Intel has a big sales team, and they will do what it takes to make the sale.

AMD's desktop CPU prices are already disruptive. Who knows what they're planning for the server parts?

 

[.vodka]

I can't be the only one noticing C15 in the name and 16 in the timings, though.

Also
http://ranker.sisoftware.net/show_r...dae7d4e3d0e2d6f082bf8fa9cca994a482f1ccfc&l=en
Looks like Zen's cache size certainly affects bandwidth, in comparison to 6950X.

Well, the results for the 6950x will be higher because of the two extra cores, and because of having to provide enough bandwidth to its load/store capabilities and its full speed AVX/2 execution engine.

Besides... that 6950x is running at 4.2GHz.


What does a stock 6900k do in this test? You know, 8C16T vs 8C16T at similar clock speeds (around 3.6GHz)

 

[itsmydamnation]

look at this guy
http://ranker.sisoftware.net/show_r...dae7d4edd5e4d6f082bf8fa9cca994a482f1ccfc&l=en

my ivb @ 4.3
is
L1D 545.09gb
L2 269.44gb
L3 157 gb

so pre thread
L1D 68.14
L2 33.68
L3 19.62

So a Zen per thread and normalized to 4.3 ( for L1D and L2 and a relative increase for L3)
L1 57.99
L2 44.05
L3 30.34

So Zen's caches really do look fixed

 

[lolfail9001]

Well, the results for the 6950x will be higher because of the two extra cores, and because of having to provide enough bandwidth to its load/store capabilities and its full speed AVX/2 execution engine.

Here is a stock 5960X if you want: http://ranker.sisoftware.net/show_r...dae7d5e3d6efdcfa88b585a3c6a39eae88fbc6f6&l=en

Yep, it is not exactly confident, but i fail to find a single haswell/broadwell result with lower l1 bandwidth.

EDIT: Upon further inspection, my belief that Zen certainly has faster L3 is reaffirmed, looks like L2s are comparable in speed.

 

[majord]

L1 is expected to be higher, it's 64b/cycle vs 32b/cyc for Zen

 

[.vodka]

L1 bandwidth is supposed to be that high, it measures >1TB/s since Haswell either here or in AIDA64's cache benchmark

Putting that aside, Zen's L2 looks to be just fine vs Haswell-E's, and its L3 is very fast for its size. It's all certainly higher than whatever any FX chip ever posted.

AMD has finally fixed their caches. Excellent!

 

[scannall]

And that's still a legally enforceable promise, not a gentleman's agreement.

Perhaps, but it is unlikely anyone would pursue a legal remedy. BUT, you would never get a sample from anyone again. Nor would your employer. If that's how you make your living... Time to learn something else.

 

[Doom2pro]

Not sure if it's been mentioned here but the ASUS X370 Crosshair VI Hero has AM3 & AM4 mounting holes.

Hopefully other boards are doing the same.

 

[lolfail9001]

Putting that aside, Zen's L2 looks to be just fine vs Haswell-E's, and its L3 is very fast for its size. It's all certainly higher than whatever any FX chip ever posted.

It looks to be faster than whatever Intel has until Skylake-X too, while we're at it.

Hey, remember Shintai going on about cache improving performance in games and not core count? Well, bad luck for him, Zen has it on both fronts.

 

[Arachnotronic]

It looks to be faster than whatever Intel has until Skylake-X too, while we're at it.

Yes, AMD appears to have designed an excellent cache subsystem for Zen.

 

[MajinCry]

It looks to be faster than whatever Intel has until Skylake-X too, while we're at it.

Hey, remember Shintai going on about cache improving performance in games and not core count? Well, bad luck for him, Zen has it on both fronts.

He's been staying away from this thread. Huh.

Wasn't he big on the "I'm not biased, AMD just can't" thing? Wonder if he's a shill.

Insulting other members is not allowed.
Markfw
Anandtech Moderator