Official AMD Ryzen Benchmarks, Reviews, Prices, and Discussion

[Doom2pro]

Looks like a bios update:
https://www.reddit.com/user/onotech


It cannot be a bad thing regarding performance or else AMD wouldn't have bothered putting all of these people thru additional labor.

I agree it has to be some significant uplift to warrant getting ALL the reviewers to rerun their tests...

 

[Magic Hate Ball]

I agree it has to be some significant uplift to warrant getting ALL the reviewers to rerun their tests...

Guys... we left debug mode on.

Oops.

 

[Minkoff]

Looks like a bios update:
https://www.reddit.com/user/onotech


It cannot be a bad thing regarding performance or else AMD wouldn't have bothered putting all of these people thru additional labor.

When you get, handcrafted wooden crates with free goodies inside, I think they can handle it. Not to mention, that their click revenue is going to explode, so I'm not really that...sympathetic

 

[MBrown]

Does the embargo/NDA lift at midnight eastern time?

 

[IEC]

Does the embargo/NDA lift at midnight eastern time?

No, 3PM GMT. So should be around 9AM Central.

Though there might be some delay in some sites' reviews if the last minute changes are true.

 

[looncraz]

Bad news?

https://twitter.com/GamersNexus/status/837026790526881797

Hope there wont be callbacks as with sandy bridge

The comment literally means that the performance can be tweaked immensely - or is quite sensitive to certain settings.

The responses are mostly doom and gloom... which just shows how little people understand - there's nothing wrong with a brand new architecture having to feel its way to its optimal settings.

 

[looncraz]

I want the board... 1800X is literally on its way to me but not a peep from Newegg on the Asus C6H. Except release date listed as... Mar 9

I ordered both from Amazon with a couple minutes of them being available. I'm probably one of the first 10 who ordered the 1800X from there (was refreshing my search every couple minutes ).

After that I went looking for motherboards. I had two to choose from - but the AsRock X370 big boy was nowhere to be found, so Asus C6H it was.

 

[Justinbaileyman]

Dont know how they could do a recall... My stuff is already in transit and will be here on Sat. Hope Fedex delivers on Saturdays. Otherwise its Monday for me.

 

[unseenmorbidity]

I wonder when tech retailers will have Ryzen cpu and boards in stock to be ordered? Newegg was the only one to get asrock boards in for preorder.

 

[DarkKnightDude]

They're bios updates. We got email notifications of them as well.

New architecture is going to take time I guess.

 

[Ajay]

I'm just glad we'll have reviews tomorrow. It was getting pretty sporty here.

 

[Rifter]

I'm just glad we'll have reviews tomorrow. It was getting pretty sporty here.

Yeah, im so ready for some real numbers. The issue is there will still be blue and red sites slanting the results one way or the other, so it will remain "sporty" for a while yet i imagine.

 

[cytg111]

have you seen the bios shots from canard? lookup their twitter.. bios has performance bias option for CB15 11.5 and more.. is that cheating or just creative tinkering?

 

[IEC]

have you seen the bios shots from canard? lookup their twitter.. bios has performance bias option for CB15 11.5 and more.. is that cheating or just creative tinkering?

ASUS has added these to Intel chipset motherboards in the past as "Xtreme Tweaking":
https://rog.asus.com/articles/maxim...ngs-for-overclocking-maximus-vi-motherboards/

There's nothing unusual about ASUS having special BIOS tweaks for benchmarking.

I imagine reviewers having received a mix of motherboards and BIOSes won't be enabling those options except for S&G.

 

[unseenmorbidity]

have you seen the bios shots from canard? lookup their twitter.. bios has performance bias option for CB15 11.5 and more.. is that cheating or just creative tinkering?

I heard that is a common feature is asus boards?

Seems like a good feature, assuming it works.

 

[formulav8]

I still have mobo's and such that says things like "Enable this for higher 3DMark 2001 benchmarks".

Is it Cheating when tweaking settings that's performs better for certain workloads? I do not think so.

 

[IEC]

Oh Linus...

https://twitter.com/LinusTech

RIP Crosshair VI Hero. Now I know why mine hasn't shipped yet

 

[Doom2pro]

Oh Linus...

https://twitter.com/LinusTech

RIP Crosshair VI Hero. Now I know why mine hasn't shipped yet

If anyone doesn't know Linus, that is a joke with him, dropping stuff... Pretty sure he is trolling everyone

 

[unseenmorbidity]

Linus destroys so much hardware that it's comical.

 

[itsmydamnation]

I am talking about die size, man. By the way, you have already ate your words on PassMark latency report, didn't you?

What words have i eaten, we dont know enough about Zen latency yet, Passmark reports my ivb random latency as 74ns. Passmarks latency single click test reports latency as only twice my L3 latency (20 something), AIDA reports it @ 50 something ( at work so i cant look). if i was to guess i would say the core or the L2 perfetcher is matching passmarks read pattern wouldn't you?

Yes latency is important and it will be dissapointing if it is real world Zen is 90ns while skylake is 50 something with comparable DDR4. if you have lots of misses and lots of stalls to memory then thats where SMT helps alleviate that problem so in most cases that extra (worst case) 35ns its not that bad.

"Let's make a small core and LARGE ARSE interconnect that takes as much as space as four cores and 8 mb of L3". That sort of philosophy?

Looks like you haven't been keeping up with issues in silicon production, WATTS per mm, power hot spots, etc
But more to the point why does it matter, the manufacture cost is so low vs the development cost and on going reoccurring business costs that its far more important that you can see at good margin which AMD can with Zen

Because i don't remember the second part. You have a point

though, i sometimes forget that Zeppelin is just like HEDT is just a server reject.

Can you just cut the crap? honestly? /Sebastian Vettle, The chips that we the highend desktop like are actually different to the chips Server market want.
We want the high clocking high leakage 4.5ghz + capable chips. 32 core Server wants the low leakage, low power in 2-3ghz target range.

So no not server rejects, your just wrong.

Obligatory reminder that these supposed "monsters" are only useful for scale-outs.

You forget that Intel salvages defective GPUs just fine as E3 Xeons.

If you call the entire enterprise and cloud and DC markets "scale out" then sure. Depending on price scaling in the line up P2 24 core a P with 16 channel memory ( allowing either 16 or 32gb dims) will be the best $ for VM density product on the market. Add in ~24 direct full bandwidth NVME drives ( its hard to fit more in 2RU) and it is also by far the best hyper converged platform as well.

 

[Genx87]

First time in a decade I am excited about an AMD CPU launch. I can't wait for tomorrow to see what they have been working on for the past 5 years.

 

[crashtech]

<snip>...The chips that we the highend desktop like are actually different to the chips Server market want.
We want the high clocking high leakage 4.5ghz + capable chips. 32 core Server wants the low leakage, low power in 2-3ghz target range.

So no not server rejects, your just wrong...<snip>

High leakage chips could be considered server rejects, in a way. Binned differently, to put it more mildly.

 

[TemjinGold]

First time in a decade I am excited about an AMD CPU launch. I can't wait for tomorrow to see what they have been working on for the past 5 years.

First time in a decade I'm excited about a CPU launch period.

 

[itsmydamnation]

High leakage chips could be considered server rejects, in a way. Binned differently, to put it more mildly.

I think the only one you can really try and claim that is the 1700 and that's assuming that it cant clock as high as 1700x or 1800x because it would be a chip that wasn't good enough for server from a clocking/power perspective and wasn't good enough for high clocks. Otherwise every chip that isn't a 7700k, 6950k or 2699 is a reject and at that point the word is kind of meaningless.

 

[formulav8]

I can't believe there are persons expecting AMD to overtake Intel on full performance. That's some wacky expectations. To truly go from basically the last in many things to the highest in one Gen with AMD's budget is quite the unrealistic expectation is it not?

AMD will end up truly competitive, NOT performance leader. Why do people keep raising the performance bar to the top?

When it comes to price/performance, then yes, AMD will be on the high end of that list. NOT outright performance king though.

I hope I'm wrong on even performance king, but I think I'm not. Especially the clock speed AMD's cpu's are shipping at.

If you simply need over 200-300fps @ 720p then buy Intel.

If you want to game with good frames, but still want to encode, broadcast, twitch, and whatever your gaming situation, in real-time at the same time, you may want to consider Ryzen.