-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
*** AMD "Zen" thread (inc AM4/APU discussion) ***
- Thread starter Boomstick777
- Start date
Short of them driving round to your house and handing you a Zen processor, how do you expect them to make it more verifiable? They have provided the files needed to do the benchmark yourself and people with 6900's have tried and found that the results match for their chips. If AMD were lying it would come out pretty quickly when people had the chips in their hands.
What you probably mean to say is not "verifiable" but "biased". In which case, I have to ask what exactly you mean by bias because they're previewing the chip with a single benchmark so what exactly would you like them to have picked that would be better? Cinebench that notoriously favours Intel? A game which would not show the CPU performance usefully? Blender is open source software and pretty platform neutral (i.e. it's not known for favouring Intel or AMD particularly). I don't get your objections.
AMD are using open source software , and they provide the demo file used - even the anti AMD hardcop have accepted AMD have done more to prove the live results are legitmate as anyone can use the same file , and direct download the same software.
they even only link to the blender website - they do not provide the software
I was wondering about this - I suppose the most extreme conspiracy theorist would say AMD could have hacked the source to perform better on Zen and so far not released it. We don't have any proof they used the released code. Unbelievable and pointless though as it'd bite them after release.
using a direct download from blender themselves and the supplied demo file , is demonstrating comparable numbers on released intel kit to the live demo.
that's enough qualitative data to support what they did is genuine
Soldato
- Joined
- 21 Jul 2005
- Posts
- 20,713
- Location
- Officially least sunny location -Ronskistats
I think you are best leaving your energy to something else and hope AMD delivers a decent product - which is the only way you will (for a very short time) silence the naysayers.
Of course some are not just naysayers but anti product X, but I am sure the regulars on here know or should know what to expect from certain folk that post.
Forget the benchmarks just wait for it to get out now it is so near.
I think you are best leaving your energy to something else and hope AMD delivers a decent product - which is the only way you will (for a very short time) silence the naysayers.
Of course some are not just naysayers but anti product X, but I am sure the regulars on here know or should know what to expect from certain folk that post.
Forget the benchmarks just wait for it to get out now it is so near.
no the thing is most of this talk is just speculation.solid benchmarks have not been shown just one slightly glanced over.its a clever tactic from amd really.
if they would have shown their full hand and in everything else its slower than intel counterparts then many would just say oooh another amd cpu failure.
by showing next to nothing everyone even if a intel fanboy is still interested in how they actually perform. so they lost next to no one from showing nothing.
so i dont think we will see anything much now solid until release.
i do hope amd can compete will get lazy intel back pushing forward again.i just hope people are realistic in their expectations and not still in dream mode when launched.
CPU Rendering scales in a near linear fashion with IPC, Cores, Threads, clocks. It is a good benchmark the same with video transcoding.
Soldato
- Joined
- 19 Feb 2011
- Posts
- 5,849
There's going to be a lot of people with egg on their faces come RyZen launch...
I put down most of the negativity from certain sections as their self justification for spending well over the odds on Intel hardware, especialy going to be worse if AMD builds are going to offer more at even cheaper price points. Resale value on a lot of Intel builds will take an absolute hammering if AMD are actually on the money with RyZen and cheaper by far.
I put down most of the negativity from certain sections as their self justification for spending well over the odds on Intel hardware, especialy going to be worse if AMD builds are going to offer more at even cheaper price points. Resale value on a lot of Intel builds will take an absolute hammering if AMD are actually on the money with RyZen and cheaper by far.
4 core 6 core and 8 core ^^^ apparently.
Shillmarks? Blender? Blender usually favours Intel, its also very widely used and recognised.
Of course that still doesn't mean its results should be accepted as indicative of its overall performance vs Intel.
Having said that for some people there is never a point where Zen would be a match or better than the 6900K, no mater which benchmarks, how many different types, by how many reviewers it will always be not true and anyone who accept's what it blatantly obvious is a shill.
Are you that guy?
Shillmarks? Blender? Blender usually favours Intel, its also very widely used and recognised.
Of course that still doesn't mean its results should be accepted as indicative of its overall performance vs Intel.
Having said that for some people there is never a point where Zen would be a match or better than the 6900K, no mater which benchmarks, how many different types, by how many reviewers it will always be not true and anyone who accept's what it blatantly obvious is a shill.
Are you that guy?
Caporegime
- Joined
- 18 Sep 2009
- Posts
- 30,291
- Location
- Dormanstown.
I think Zen will be good, I've said nothing contrary to that.
But if Cinebench is so "bad" and "Intel biased", then why is it a 4670K is faster than an FX83 in Blender (Baring in mind Blender's being touted as some brilliant benchmark) but the opposite is shown in Cinebench (As you'd expect). This is something that would *never* be accepted because it's an outlier. An FX83 going 100% versus an i5 4670K is faster than the i5, not slower.
I understand people being hyped and positive for Zen (As I am) but it suddenly seems like because a certain benchmark supports an agenda, it's fair game.
I'm not saying the results are faked or anything like that, I just have absolutely no faith in a benchmark which posts outliers as per the FX83 versus i5.
I'm not shocked AMD released the wrong benchmark files originally either, but I've got no conspiracy for that. It's not an AMD product unless there's a mess up along the way.
But if Cinebench is so "bad" and "Intel biased", then why is it a 4670K is faster than an FX83 in Blender (Baring in mind Blender's being touted as some brilliant benchmark) but the opposite is shown in Cinebench (As you'd expect). This is something that would *never* be accepted because it's an outlier. An FX83 going 100% versus an i5 4670K is faster than the i5, not slower.
I understand people being hyped and positive for Zen (As I am) but it suddenly seems like because a certain benchmark supports an agenda, it's fair game.
I'm not saying the results are faked or anything like that, I just have absolutely no faith in a benchmark which posts outliers as per the FX83 versus i5.
I'm not shocked AMD released the wrong benchmark files originally either, but I've got no conspiracy for that. It's not an AMD product unless there's a mess up along the way.
the real thing is price for the performance.if its priced well and performs well its going to be great.
no sides can really argue about anything as next to nothing is known.we got till what january . so thats gunna be a hell of a long time with nothing but speculation.
no sides can really argue about anything as next to nothing is known.we got till what january . so thats gunna be a hell of a long time with nothing but speculation.
Caporegime
- Joined
- 18 Sep 2009
- Posts
- 30,291
- Location
- Dormanstown.
Pricing will be key for Zen.
Lisa Su's on record not wanting AMD to be the budget brand, and if they've got the performance, there's no reason why they need to be.
I don't suggest Fury X pricing (Pricing above your competitor while being inferior). But there's no reason why they have to be dirt cheap either.
Of course because of the pound, there has to be some forgiveness of UK pricing.
Lisa Su's on record not wanting AMD to be the budget brand, and if they've got the performance, there's no reason why they need to be.
I don't suggest Fury X pricing (Pricing above your competitor while being inferior). But there's no reason why they have to be dirt cheap either.
Of course because of the pound, there has to be some forgiveness of UK pricing.
Last edited:
The idea that CineBench is skewed toward Intel comes from the discovery of code in it which throttles performance if an AMD CPU is detected, but that was some years ago and a couple of versions back, i'm not convinced its still happening.
There is a disparity between CineBench and and Blender despite them performing the same task, i believe this is because CinBench is natively an X86 application, Blender is not, its Unix Based and ported to X86, i'm sure that's the problem with it on Windows, certainly its a lot faster on Linux.
There is a disparity between CineBench and and Blender despite them performing the same task, i believe this is because CinBench is natively an X86 application, Blender is not, its Unix Based and ported to X86, i'm sure that's the problem with it on Windows, certainly its a lot faster on Linux.
Having a think about it, the disparity between Cinebench and Blender is more than likely down to what each of them is using within the CPU.
Blender is float heavy with its 3D rendering so it will be stressing the FPU's in the CPU's the most. But having a read of Cinebench it says that it uses, quoted 'various algorithms to stress all available processor cores.'
So with cinebench it could be stressing Float and Int performance as it says it is a full CPU bechmark, where as Blender is Float heavy and purely geared towards 3D rendering. This would show up as a disparity between the i5 4670k and 8350 in blender compared to cine-bench becasue of Bulldozers poor FPU performance.
I don't know the entire in's and outs of both programs but it seems reasonable that the above points to the disparity.
So with Blender they were showing the greatly improved FPU performance in Zen, while with Handbrake it was the greatly improved Integer performance. Showing that AMD with Zen have brought both Integer and Float performance up to Intel levels.
Blender is float heavy with its 3D rendering so it will be stressing the FPU's in the CPU's the most. But having a read of Cinebench it says that it uses, quoted 'various algorithms to stress all available processor cores.'
So with cinebench it could be stressing Float and Int performance as it says it is a full CPU bechmark, where as Blender is Float heavy and purely geared towards 3D rendering. This would show up as a disparity between the i5 4670k and 8350 in blender compared to cine-bench becasue of Bulldozers poor FPU performance.
I don't know the entire in's and outs of both programs but it seems reasonable that the above points to the disparity.
So with Blender they were showing the greatly improved FPU performance in Zen, while with Handbrake it was the greatly improved Integer performance. Showing that AMD with Zen have brought both Integer and Float performance up to Intel levels.