In leaked benchmarks, AMD Zen falls short of Intel Haswell, but shows promise nonetheless

[msx68k]

Hey, Shawn! Are you an Intel fanboy? I don't think so. Please, read carefully the result of that benchmark: The I7@3.6GHz is just 24% faster than I5@3.4GHz.
Is possible that an Intel 8 threads CPU be just 24% faster than a 4 threads one that run at lower speed? Obviously not. The difference would be arround 100%.
The answer is very clear: That test was made with the 4 processors using just four threads. (Furthermore, it is the most honest way to do the test, and see the real performance difference between cpu.)
In this scenario, and considering the clock speed, the Zen has the highest FPS/GHz rate of all. The Zen performance is close to the Skylake.

 

[Adam Leite]

I am a big fan of AMD. Have been since 1998, when I overcame my fear and assembled my first AMD. I don't play games but assemble computers for gamers. It is hard to convince them, but once you do the first, the rest comes along. Most still prefer Intel variations, but not all of them can pay the price.

 

[mlscrow]

AMD fans are rejoicing in the street, because there is something to rejoice about. Yes, a 2.8GHz Engineering sample bests an i5 haswell K edition clocked at 3.4Ghz. This is incredible news and performance. You can talk all you like about "Oh, an AMD part performs as good as 3 year old Intel tech"...actually, it performs as good as today's Intel tech, considering today's Intel tech isn't all that much better than 3 years ago, durp.

Second, not only does it outperform Intel parts, it does so at much lower frequency and with an Engineering Sample at that. Everyone knows that ES's are a shadow of what their production counterpart is capable of. This just means that we're going to see an outstanding product when it is released toward the end of this year, and that's 2016 everyone. Masses will have it in 2017, but we'll have something in 2016. This was spoken by Lisa Su. You want links, Google them, or simply pay attention to market news as it happens, don't make other people do it for you simply because you're lazy. =P

 

[namesrejected]

Steve do you feel the love?

haha yes I do, you guys rock

Can I get a free shirt if I say " your the best " too ?

 

[namesrejected]

Urg! I read about a 1/4 through all this and gave up... We will see when ZEN gets here, hopefully its a good overclocker. That's all I have to say.

 

[poohbear]

If it nearly matches a 4790 at only 2.8ghz, how on earth is this a disappointment?!? AMD is NOT trying to win the performance crown, they're trying to provide awesome performance at a decent pricepoint, this leaks proves that! Who would pay $300+ for a 4790 when they can get a similarly performing CPU @ $200??

 

[Squid Surprise]

If it nearly matches a 4790 at only 2.8ghz, how on earth is this a disappointment?!? AMD is NOT trying to win the performance crown, they're trying to provide awesome performance at a decent pricepoint, this leaks proves that! Who would pay $300+ for a 4790 when they can get a similarly performing CPU @ $200??

Wow, AMD has already stated the Zen would be $200? Must have missed that announcement.... Can you provide a link, I can't find one

 

[fps4ever]

I'm sorry but it disappoints me even more knowing the Tech Editor is backing an author with an obvious negative opinion on these benchmarks, even though the numbers at worse tell them it's at the very least competitive. Try looking around the web at other news reports of these numbers along with comments. Nobody but you guys are reporting this as "it's not a great start for Zen". There are a lot of knowledgeable people around the web like myself that know a thing or two about these technologies as well. Yes by all means I agree, why don't we wait for more benchmarks before writing that in an official TS article.

 

[Steve]

I'm sorry but it disappoints me even more knowing the Tech Editor is backing an author with an obvious negative opinion on these benchmarks, even though the numbers at worse tell them it's at the very least competitive. Try looking around the web at other news reports of these numbers along with comments. Nobody but you guys are reporting this as "it's not a great start for Zen". There are a lot of knowledgeable people around the web like myself that know a thing or two about these technologies as well. Yes by all means I agree, why don't we wait for more benchmarks before writing that in an official TS article.

I am assuming the numbers are fake given the source and feel they aren't worth arguing over. I was simply stating that I don't feel Shawn's comments were out of line. He did a pretty good job of listing the cons and pros but ultimately it is one test, using a suspect hardware configuration. If you are disappointed with anything I have said in this thread then you for sure shouldn't read my Zen coverage when it is finally available. I will horrify you by trying so stick with the facts and figures.

 

[misor]

Only a hardcore fanboy would believe everything amd say or claim. take a look at previous cpu gen releases.
if it so happen that the zen cpu engineering sample is a good representation of the consumer edition, good things will come for amd. but as they say, when someone is claiming heaven and earth, always take it with a grain of salt.

I think tech news editors should be as objective as possible. if they were fanboys to a fault and were proven to be wrong on judgment day (zen consumer release), then they lose credibility on that day but never today.

 

[Ryrynz]

I still consider my 3770k high end. To this day I still cannot justify the cost of a CPU upgrade and it's going on 4 years. What happened to the days of 4 year old tech being obsolete?

"Consider it" all you like, but It's not. Six and eight cores demolish it in various multithreaded tasks, The 6700K is on average about 20% faster but more so in some tasks. It may be suitable for your purpose but it's not high end.

 

[Ryrynz]

"The 6700K is on average about 20% faster but more so in some tasks. It may be suitable for your purpose but it's not high end.

Probably more like upwards of 25% faster really. Also the integrated graphics is significantly faster but I guess that'snot cared about. Intel is more focused on power savings, die size and GPU performance than x86 performance because let's face it, it has zero competition and doesn't need a a non GPU quad core with 256MB EDRAM. But if it wanted to it sure could, and then you'd see an instant ~20% performance increase from that alone. We won't see anything good from Intel until 2018 when their Icelake processors come out which is an entirely new design. Zen is catch up.. it won't make Intel step up it's x86 game, not yet.

 

[yRaz]

"Consider it" all you like, but It's not. Six and eight cores demolish it in various multithreaded tasks, The 6700K is on average about 20% faster but more so in some tasks. It may be suitable for your purpose but it's not high end.

Maybe 20% faster in synthetic benchmarks, Most I've really been able to find has been in the 12-15% with things like gaming being in the single digits. but if I ever feel the need I can just slap on a 30% overclock and have performance on par. Also, 20% isn't worth $400 to me and if it is to you then you're a fool.

 

[hood6558]

"The Zen ES reports eight cores, 16 threads and is a 14nm part yet it can’t keep pace with Intel’s old chips?"

"Making matters even worse for AMD is the fact that Intel is planning to release Cannonlake, its 10nm die shrink of Kaby Lake, next year. AMD has said it doesn’t expect to ship Zen with full market availability until sometime in 2017."

I kind of agree with HardReset because of the statements made above in the article. HardReset makes some good points and I have a few extra thoughts of my own.

If the 8-cores aren't being fully utilized than making that comparison statement seems unjustified. But the ambiguity in the statement "Intel's old chips" is not helping the matter either.

Intel is always planning things but the plans don't always go exactly as...planned. One thing that is squarely on AMDs side is the apparent diminishing returns on Moore's Law. Intel's gains may become much less and on a less frequent basis than AMDs by the simple fact that transistor counts aren't doubling every 2-years. Intel is driving on unpaved dirt roads now while AMD is still driving on highways. They may be behind but they aren't out of the race.

So while it's easy to say things look bad for AMD, if they are making gains at a more rapid pace than Intel, even though they are behind right now, I don't see how that can be viewed as a bad thing.

And lastly the statements above completely disregard bang for buck, integrated APU performance, and TDP consumption.

Very sensible viewpoint. Also, we don't know how Zen will overclock - it could be a question of, will the cheaper price be negated by the need for more expensive cooling to get high clocks (as in the past). I think people hoped Zen would be another Athlon 64 (when AMD was faster than Intel in 2003).(http://www.pcmax.com/amd_intel.html)

 

[Squid Surprise]

The thing is, most of us (not including Intel and Nvidia employees) want AMD to succeed. Monopolies are not good for the consumer, and as long as AMD can't compete, we'll be paying top dollar for incremental improvements.

But... Just because we WANT AMD to succeed doesn't mean they will! No one can predict the future - but the best way to try is to base it on past performance... And AMD hasn't competed for over a decade.... Let's hope they do - but nothing we've seen here has shown that they will.

 

[ScubaRhys]

So this thing is slower than the CPU I bought 2 years ago? Yeah, I won't be upgrading. Can't imagine many will tbh.

 

[fps4ever]

I'm sorry but it disappoints me even more knowing the Tech Editor is backing an author with an obvious negative opinion on these benchmarks, even though the numbers at worse tell them it's at the very least competitive. Try looking around the web at other news reports of these numbers along with comments. Nobody but you guys are reporting this as "it's not a great start for Zen". There are a lot of knowledgeable people around the web like myself that know a thing or two about these technologies as well. Yes by all means I agree, why don't we wait for more benchmarks before writing that in an official TS article.

I am assuming the numbers are fake given the source and feel they aren't worth arguing over. I was simply stating that I don't feel Shawn's comments were out of line. He did a pretty good job of listing the cons and pros but ultimately it is one test, using a suspect hardware configuration. If you are disappointed with anything I have said in this thread then you for sure shouldn't read my Zen coverage when it is finally available. I will horrify you by trying so stick with the facts and figures.

Now you're suggesting the benchmarks are fake? I applaud you for sticking up for an author that comes up with a "not looking good for Zen" opinion based off of suspected fake benchmark numbers. I agree, I won't be reading your article on Zen when it comes out. There are plenty of other websites for that. Just want to clarify that this individual article could have been fixed with very a couple minor tweaks and am not picking on the entire website. You're right, it's really not worth arguing over. I like a lot of articles on this website, so kudos for that!

 

[cliffordcooley]

I can't believe all the judgemental comments from an engineering sample.

 

[yRaz]

I can't believe all the judgemental comments from an engineering sample.

Fanboyism knows no bounds

 

[HardReset]

I am assuming the numbers are fake given the source and feel they aren't worth arguing over. I was simply stating that I don't feel Shawn's comments were out of line. He did a pretty good job of listing the cons and pros but ultimately it is one test, using a suspect hardware configuration. If you are disappointed with anything I have said in this thread then you for sure shouldn't read my Zen coverage when it is finally available. I will horrify you by trying so stick with the facts and figures.

Taking Steve's word for "pretty good job of listing the cons and pros". When evaluating CPU's pros and
cons, following should NOT be considered:

- IPC (and everything that follows from that)
- Core count and clock speed (as high core count=lower clock speed generally)
- How many cores particular workload is using (low core usage usually means low core parts are better)
- TDP/power consumption
- Motherboard price (estimated if necessary)

 

[yRaz]

Taking Steve's word for "pretty good job of listing the cons and pros". When evaluating CPU's pros and
cons, following should NOT be considered:

- IPC (and everything that follows from that)
- Core count and clock speed (as high core count=lower clock speed generally)
- How many cores particular workload is using (low core usage usually means low core parts are better)
- TDP/power consumption
- Motherboard price (estimated if necessary)

As much as I like detailed comparisons, you cannot get accurate information like that from one benchmark of a review sample and trying to do so is irresponsible reporting, but thank you for the early analysis.

 

[HardReset]

As much as I like detailed comparisons, you cannot get accurate information like that from one benchmark of a review sample and trying to do so is irresponsible reporting, but thank you for the early analysis.

Exactly. Better choice is that no conclusions are not drawn from unverified benchmarks. But straight from article:

Nothing is official at this hour but if the benchmarks are indeed legitimate, it's not a great start for Zen.

So here's the problem. Author is indeed making analysis from those benchmarks, assuming they are legitimate. That "analysis" is missing all points I listed.

(while motherboard price is unknown, manufacturing costs can be estimated)

 

[yRaz]

Exactly. Better choice is that no conclusions are not drawn from unverified benchmarks. But straight from article:



So here's the problem. Author is indeed making analysis from those benchmarks, assuming they are legitimate. That "analysis" is missing all points I listed.

(while motherboard price is unknown, manufacturing costs can be estimated)

Just look at the title of the article. He showed skceptisim through the whole article. A detailed analysis would have been a bad idea because of the limited information provided. But with the limited information provided, I find his analysis to be accurate. There is very little to gain from over-analyizing the information provided.

 

[cliffordcooley]

Give it a rest HardReset! Yes judging by this one benchmark the author said nothing at all to provoke your anger. The author kept it short because it was not substantiating proof. Yet you are trying to rip him a new one for not holding an opinion for which he did (If you would just read).

 

[HardReset]

Just look at the title of the article. He showed skceptisim through the whole article. A detailed analysis would have been a bad idea because of the limited information provided. But with the limited information provided, I find his analysis to be accurate. There is very little to gain from over-analyizing the information provided.

I already provided much more detailed analysis. Author says:

Nothing is official at this hour but if the benchmarks are indeed legitimate, it's not a great start for Zen.

So author is strictly saying that 58% increase on IPC compared to AMD's last generation part is "not great start".

"(y)"

Give it a rest HardReset! Yes judging by this one benchmark the author said nothing at all to provoke your anger. The author kept it short because it was not substantiating proof. Yet you are trying to rip him a new one for not holding an opinion for which he did (If you would just read).

Once again:

Nothing is official at this hour but if the benchmarks are indeed legitimate, it's not a great start for Zen.

Author clearly states that IF these benchmarks are accurate, THEN it's not great start to Zen.

Now, IF 58% increase on IPC over last generation is not great start, THEN what is?