Zen is faster than Broadwell-E clock for clock

[Sintezza]

8 minutes ago, i_build_nanosuits said:

From what we've seen so far Zen seem to be a lot better...if they can deliver this product in a slightly faster clocked quad-core flavor for a little less money than the intel counterpart this could be very interesting...it's just sad that it took so long, now intel has had time to increase the average price of the CPU's and now AMD will just price it accordingly in regards to performance and intel's pricing...

Intel had kept the same pricing for the last 3 or 4 generations prior to haswell...but the latest stuff skylake and broadwell has seen prices increase...so it took amd just a couple years too long.

 

At the time they will release it, they will also be 2 gens behind again basicly, except for the enthusiast line.

However the blender test showed in that AMD video doesnt make any sense to me.

Honnestly it doesnt realy proof anything as long as such kind of tests arent done by third party unbiassed testers.

According to gaming, the earlier Ashes of the Singularity leaks, a game that scales realy well upto 8 threads, it didnt realy look that great for zen yet.

But again, that are also sources that cannot be vallidated.

 

 

[Qub3d]

Mmm... The disagreement in this thread is feeding my desire for sodium.

[i_build_nanosuits]

1 minute ago, Sintezza said:

 

At the time they will release it, they will also be 2 gens behind again basicly, except for the enthusiast line.

However the blender test showed in that AMD video doesnt make any sense to me.

Honnestly it doesnt realy proof anything.

no it doesnt, anything coming straight from the company like that is to be taken with a mountain of salt...they probably used slower RAM configuration and slower storage device on the intel platform to slow it down a bit, and probably used the best possible stuff on the AMD platform to get it to perform as best as possible...stuff like that...and they probably picked this blender test because it's a best case scenario where the CPU react better to the software and intel is doing not so well for example etc.

 

it's 100% sure that they cherry picked the components and the testing in order to make their cpu shine...but still, it's not bad...if that is 3.0 ghz on both of them...oh and the intel part can also have the on-cache running at a very slow speed etc...let's just say there are multiple ways to slow down a CPU while having it running at 3.0ghz...thermal trottling is another...not saying they did just that...but a combination of some of those IMHO is very likely.

[byalexandr]

23 hours ago, iDeFecZx said:

So for the AMD CPU to be as fast, the Intel one need to be handicapped?

They put both at 3.0GHz to do a comparison of the IPC. AMD mentioned that Zen would be well above 3.0GHz, just as Intel's Broadwell chips are.

[i_build_nanosuits]

1 minute ago, byalexandr said:

They put both at 3.0GHz to do a comparison of the IPC. AMD mentioned that Zen would be well above 3.0GHz, just as Intel's Broadwell chips are.

read the post i did just above yours...

[Sintezza]

6 minutes ago, i_build_nanosuits said:

no it doesnt, anything coming straight from the company like that is to be taken with a mountain of salt...they probably used slower RAM configuration and slower storage device on the intel platform to slow it down a bit, and probably used the best possible stuff on the AMD platform to get it to perform as best as possible...stuff like that...and they probably picked this blender test because it's a best case scenario where the CPU react better to the software and intel is doing not so well for example etc.

 

it's 100% sure that they cherry picked the components and the testing in order to make their cpu shine...but still, it's not bad...if that is 3.0 ghz on both of them...oh and the intel part can also have the on-cache running at a very slow speed etc...let's just say there are multiple ways to slow down a CPU while having it running at 3.0ghz...thermal trottling is another...not saying they did just that...but a combination of some of those IMHO is very likely.

If they wanne do something unbiassed, then they should have shown something like super pi calculations single and multithreaded.

That would give a better overall image.

Next to that they should have used multiple render applications aswell.

[i_build_nanosuits]

1 minute ago, Sintezza said:

If they wanne do something unbiassed, then they should have shown something like super pi calculations single and multithreaded.

That would give a better overall image.

and prove that both CPU's were running at the same clockspeeds, show the thermals, on-board CPU cache clocks, memory clockspeeds, shown that both used equivalent storage devices etc...

[Remixt]

I won't believe anything AMD INTEL or NVIDIA says. I'm waiting for reviews. Look at the crap they've pulled in the past. 

[Bhav]

Guys I just put a potato inside my PC and its faster than a 6950k with SLI Pascal Titans.

[Remixt]

2 hours ago, Bhav said:

Guys I just put a potato inside my PC and its faster than a 6950k with SLI Pascal Titans.

Careful when overclocking I've heard you can get a really bad case of hot potato. It may be better for you to use french fry coolant(sunflower oil), I've heard it's a lot more resistant to high temps.

[Matthewv1998]

On August 18, 2016 at 6:30 AM, iDeFecZx said:

So for the AMD CPU to be as fast, the Intel one need to be handicapped?

It is only an engineering sample, not even a full product yet 

[Dabombinable]

11 hours ago, Sintezza said:

Yeah the CMT design was exaly brilliant for the server market.

But for the home desktop user not so much, since software that the general home user would use wasnt optimized.

And most likely never will be.

 

Considering Microsoft had to actually release a patch to get AMD's CMT line even working as they should....

[JoeCoke]

The amount of stupid is blowing my mind. The benchmarks were for comparing IPC, so of course the intel CPU was underclocked. That Zen CPU is not final, it's a engineering sample. Are you people just THAT desperate to rip on AMD?

 

If Zen comes out and is actually good, it'll be really good for the market. Prices will go down and technology will improve. So quit being so damn biased and actually wait for real benchmarks. Fanboyism only slows progress.

[dalekphalm]

Just now, 2Buck said:

The amount of stupid is blowing my mind. The benchmarks were for comparing IPC, so of course the intel CPU was underclocked. That Zen CPU is not final, it's a engineering sample. Are you people just THAT desperate to rip on AMD?

 

If Zen comes out and is actually good, it'll be really good for the market. Prices will go down and technology will improve. So quit being so damn biased and actually wait for real benchmarks. Fanboyism only slows progress.

Yeah a lot of people just love to shit on AMD, and couldn't possibly believe that AMD could build a CPU as good as an Intel one, so they find excuses to shit on this news (They had to, *GASP*, Handicap the Intel CPU?????)

 

The way I see it is like this: This is great news so far. Still too early to know for certain, but all things are looking good. Wait for the benchmarks and reviews, and then shit on AMD if they failed.

 

But complaining about the clock speeds in an IPC comparison is rather silly. Furthermore, the amount of users who just straight up believe "Yep, the only way AMD could do this is if they sabotaged the Intel rig" is kind of hilarious.

 

C'mon guys. AMD has done some stupid shit now and then, but for them to blatantly just rig the test that badly? Nope. I don't believe it.

 

I'll personally patiently wait and see for reviews, while other people shit on AMD for no reason.

 

Of course, there's the flip side too, where people take this as evidence that AMD is gonna build a CPU that totally destroys everything Intel has to offer, and that they'll price it at sub $300 (because that makes a lot of fucking sense! ... not)

[JoeCoke]

Just now, dalekphalm said:

Yeah a lot of people just love to shit on AMD, and couldn't possibly believe that AMD could build a CPU as good as an Intel one, so they find excuses to shit on this news (They had to, *GASP*, Handicap the Intel CPU?????)

 

The way I see it is like this: This is great news so far. Still too early to know for certain, but all things are looking good. Wait for the benchmarks and reviews, and then shit on AMD if they failed.

 

But complaining about the clock speeds in an IPC comparison is rather silly. Furthermore, the amount of users who just straight up believe "Yep, the only way AMD could do this is if they sabotaged the Intel rig" is kind of hilarious.

 

C'mon guys. AMD has done some stupid shit now and then, but for them to blatantly just rig the test that badly? Nope. I don't believe it.

 

I'll personally patiently wait and see for reviews, while other people shit on AMD for no reason.

 

Of course, there's the flip side too, where people take this as evidence that AMD is gonna build a CPU that totally destroys everything Intel has to offer, and that they'll price it at sub $300 (because that makes a lot of fucking sense! ... not)

I think these people need a lesson on what IPC means. People actually think IPC lowers when you underclock. smh

 

And the AMD hate on these forums is just insane. Sure, the FX series is horrible (I know first hand), but the things I see on here blows my mind. I've seen fanboys say many times on this forum that the Athlon 64 X2 was worse than a Pentium 4 and AMD has always sucked. Yeah.... That proves that most of these kids just got started in the last few years.

 

But don't get me wrong, fanboyism from both sides is just pathetic. You should never be loyal to just one brand. Buy the better product for the price, regardless of brand. Competition is a beautiful thing, so even if you're an Intel fanboy, you should be happy for Zen, it could potentially lower CPU prices and speed up progress (If AMD truly has their heads outta there asses this time).

[Briggsy]

I would like to see these CPU's compared with Hyper-threading/SMT disabled. I hope Zen will be my next CPU upgrade, but any such weaknesses that SMT is covering up may become obvious in real world performance and I want to know about any weaknesses long before then.

[dalekphalm]

2 minutes ago, Briggsy said:

I would like to see these CPU's compared with Hyper-threading/SMT disabled. I hope Zen will be my next CPU upgrade, but any such weaknesses that SMT is covering up may become obvious in real world performance and I want to know about any weaknesses long before then.

I doubt we'll see any SMT disabled tests directly from AMD, but hopefully reviewers will benchmark the chips in every way imaginable.

[rockon5622]

Isnt zen targeted at the server side of things?

[Jumper118]

3 minutes ago, rockon5622 said:

Isnt zen targeted at the server side of things?

no

[rockon5622]

Just now, Jumper118 said:

no

How do you know? Just curious. 

[Bhav]

What I actually find astounding about this is that (so far) the OP hasnt had any likes / thumbs up etc.

 

I know a lot of other forums (cough *OCN* cough), where AMD fans would be all over this news like flies on a turd.

 

But from what we can see, its only a single benchmark (from what I heard, I havnt even bothered checking), and the Intel chip was handicapped.

[Bhav]

Just now, rockon5622 said:

How do you know? Just curious. 

They've been marketing it as the second coming of Jesus, well, as the next best, fastest, better than Intel chip.

 

They wouldnt be marketing a server chip like this.

[Jumper118]

2 minutes ago, rockon5622 said:

How do you know? Just curious. 

does have quad channel ram support. 

[rockon5622]

Just now, Bhav said:

They've been marketing it as the second coming of Jesus, well, as the next best, fastest, better than Intel chip.

 

They wouldnt be marketing a server chip like this.

Call it a hunch but the first wave of zen might be going after the servers market. I mean people in that area are probably looking for an alternative to high priced xeons. And they did put it up against an enthusiast broadwell instead of mainstream skylake.

[Bhav]

Broadwell-E isnt exactly a server chip. Xeons are.