ZEN ES Benchmark from french hardware Magazine

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

bjt2

Senior member
Sep 11, 2016
784
180
86
168% divided by 16 threads = 10.5
105% divided by 8 thread = 13.1
So, per thread Zen is slower than fx8370 in those multithreaded apps.
Taking about 20% frequency difference it make Zen thread about equal to Piledriver in Instructions per clock?

Ofcourse, if we assume fx 8370 is an 8 core CPU and that zen CPU is also 8 core CPU, Zen is a lot faster, like 70% faster
FX8370 has about 700MHz higher frequency (3.3 mean vs 4 mean)...
 

bjt2

Senior member
Sep 11, 2016
784
180
86
What I find most funny is that when I posted the power delta comparison (idle-load) of the AMD vs. Intel CPUs, which gave a delta of around 90W vs. the system power consumption of almost 200W (load), I got literally ripped into pieces (e.g. https://forums.anandtech.com/threads/summit-ridge-zen-benchmarks.2482739/page-120#post-38630370).

But now these guys have measured the exact CPU power consumption... and it exactly matches what I found by measuring the delta (93W vs. 94W).

2001347995.png


https://forums.anandtech.com/threads/summit-ridge-zen-benchmarks.2482739/page-118#post-38630275

This is the 12V line power, but do you know? There are these little tiny chips involved, called VRMs, that have at most 90% efficiency, because the CPU is not loaded with 12V but about 1.1V (probably)...
 
  • Like
Reactions: inf64

beginner99

Diamond Member
Jun 2, 2009
5,208
1,580
136
Price will be important part of the equation. If you need 7700K at ~4.8 GHz, it will cost you ~$550 ($150 for MB + $350 for CPU + $50 for cooling). On the other hand, if 4c/8t Ryzen manages to boost up to 4.2 GHz and costs less than $250, and you buy decent AM4 board for ~$80-90 and use Box(Wraith) or cheap ($20) aftermarket cooler, you could save up to $200 if you go for AM4. And that's the difference between 1070 & 1080 or 1080 and 1080 Ti.

The cooler can also be reused from previous build. A big slap of metal doesn't really age or turn bad. And don't expect to get a good OC an a $80 Mobo...
 

ultima_trev

Member
Nov 4, 2015
148
66
66
The fact it only performs 11% behind the 6900K in games with a gimped clock paints a good picture. Hopefully the 4C/8T version can OC to 4.3GHz or so on air, it will probably do very well against the 6700K / 4790K.
 

cytg111

Lifer
Mar 17, 2008
23,049
12,717
136
What I find most funny is that when I posted the power delta comparison (idle-load) of the AMD vs. Intel CPUs, which gave a delta of around 90W vs. the system power consumption of almost 200W (load), I got literally ripped into pieces (e.g. https://forums.anandtech.com/threads/summit-ridge-zen-benchmarks.2482739/page-120#post-38630370).

But now these guys have measured the exact CPU power consumption... and it exactly matches what I found by measuring the delta (93W vs. 94W).

2001347995.png


https://forums.anandtech.com/threads/summit-ridge-zen-benchmarks.2482739/page-118#post-38630275

Lots of things can still be optimized to bring wattage down I am sure, we havent seen the final product yet.
But overall Zen is looking very promising so far. Dont you agree?
 

Erenhardt

Diamond Member
Dec 1, 2012
3,251
105
101
What I find most funny is that when I posted the power delta comparison (idle-load) of the AMD vs. Intel CPUs, which gave a delta of around 90W vs. the system power consumption of almost 200W (load), I got literally ripped into pieces (e.g. https://forums.anandtech.com/threads/summit-ridge-zen-benchmarks.2482739/page-120#post-38630370).

But now these guys have measured the exact CPU power consumption... and it exactly matches what I found by measuring the delta (93W vs. 94W).

2001347995.png


https://forums.anandtech.com/threads/summit-ridge-zen-benchmarks.2482739/page-118#post-38630275

No. This is bad methodology.
Disable cool and quiet for fx8370 and its power delta will be around 50 watts, which would mean it is the most efficient CPU, which obviously is not true.

Energy consumed for a given task is what we should be looking at. End of story. Sadly, almost no review measures it.
 

jihe

Senior member
Nov 6, 2009
747
97
91
slower then an i5 6600 on games, good job amd.

Yes i understand the frequency is lower, is unoptimized or whatever your going to say guys but theres is no excuse at this point...it is clear as hell that this chip wont be the gamers choice unless they price it lower then $200 which is unlikely..

Might be good for other tasks tho..

I mean 6900K got trounced by 6700K, good job Intel?
 

coercitiv

Diamond Member
Jan 24, 2014
6,151
11,686
136
What I find most funny is that when I posted the power delta comparison (idle-load) of the AMD vs. Intel CPUs, which gave a delta of around 90W vs. the system power consumption of almost 200W (load), I got literally ripped into pieces (e.g. https://forums.anandtech.com/threads/summit-ridge-zen-benchmarks.2482739/page-120#post-38630370).

But now these guys have measured the exact CPU power consumption... and it exactly matches what I found by measuring the delta (93W vs. 94W).
Let's look at your assessment:
Zen delta: 94W
Intel delta: 85W

Now theirs:
Zen power: 93W
Intel power: 96W
Do you still consider their data exactly matches your data?
 
Aug 11, 2008
10,451
642
126
The fact it only performs 11% behind the 6900K in games with a gimped clock paints a good picture. Hopefully the 4C/8T version can OC to 4.3GHz or so on air, it will probably do very well against the 6700K / 4790K.
OTOH, it performs nearly 10% *worse* in games than the hex core HEDT chip, 30% worse than a 6700k, and no better than an i5. If one accepts the data from this review, I think the power consumption looks excellent, the productivity benchmarks are pretty good (in line with what I expected), performance per watt in productivity tasks is very good, while I dont really understand how anyone can look at the gaming benchmarks and think they are good news. Granted, the clockspeeds are low, but the hex core intel has at least 20% more clockspeed headroom as well.
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
The cooler can also be reused from previous build. A big slap of metal doesn't really age or turn bad. And don't expect to get a good OC an a $80 Mobo...

Existing (Socket 754 - FM2+) coolers can be only used if it attaches to the original motherboard mounting brackets. If the cooler has it's own mounting kit, then it is incompatible with AM4 without a new AM4 specific mounting kit.

19123950603l.jpg
 

witeken

Diamond Member
Dec 25, 2013
3,899
193
106
Let's look at your assessment:

Now theirs:

Do you still consider their data exactly matches your data?
The Zen power is equal. The Intel power is not, but certainly not too bad. I never said it would match exactly, just that it gives a better proxy than the raw system power. (Approximation of 85W vs. 96W is much better than 200W vs. 96W.) Power consumption of course can also vary because of different workloads, cooling that allows for greater frequencies, silicon lottery and different idle power, which for HEDT chips I would place on the order of plus minus 10W.
 

inf64

Diamond Member
Mar 11, 2011
3,685
3,957
136
.. I dont really understand how anyone can look at the gaming benchmarks and think they are good news. Granted, the clockspeeds are low, but the hex core intel has at least 20% more clockspeed headroom as well.
Because, drum roll, it is an ES with very low clock speeds and malfunctioning Turbo feature? CanardPC author said that even tho the label on the ES suggested ST Turbo of 3.5Ghz, their ES never went over 3.4Ghz and usually sat between base 3.15Ghz and 3.3Ghz. Retail SKUs won't have such "feature", on the contrary. Almost all games they tested rarely tax more than 4 cores so the ES in question did very well, scoring 10% lower than intel parts that clock north of 3.7Ghz in games.
 
Aug 11, 2008
10,451
642
126
Because, drum roll, it is an ES with very low clock speeds and malfunctioning Turbo feature? CanardPC author said that even tho the label on the ES suggested ST Turbo of 3.5Ghz, their ES never went over 3.4Ghz and usually sat between base 3.15Ghz and 3.3Ghz. Retail SKUs won't have such "feature", on the contrary. Almost all games they tested rarely tax more than 4 cores so the ES in question did very well, scoring 10% lower than intel parts that clock north of 3.7Ghz in games.
Well, drumroll, personally, I expect intel to maintain or even extend the clockspeed advantage with max overclocks.
 

vissarix

Senior member
Jun 12, 2015
297
96
101
I highly doubt Core i5 Kabylake to be better in 2016/2017 games than any 6-8 Core Intel/AMD ZEN CPUs. Gamers should understand by now that fps is not the only metric for better gaming, we should also take in to consideration the FrameTimes.
Your words sounds familiar...lol...hmm let me think when ive heard them before...oh yeah...bulldozer release lmao...yeah but when software will use more cores bulldozer will stomp everything...that was 5 years ago...and here we are...almost 2017 and an i3 humiliates bulldozer...

Same story repeats himself over an over again...
 
  • Like
Reactions: Sweepr

inf64

Diamond Member
Mar 11, 2011
3,685
3,957
136
Well, drumroll, personally, I expect intel to maintain or even extend the clockspeed advantage with max overclocks.
Oh I have no doubt that intel parts will have the max OCing edge over Zen. The problem will be that you will have a choice between similarly priced 8C/16T Zen clocking to ~4-4.3Ghz Vs 4C/8T 7700K clocking to 4.6-4.8Ghz. Or even worse, 2x cheaper 4C/8T Zen clocking to 4-4.3Ghz VS 7700K. In games the 4C/8T highly clocked Zen will likely perform better than 8C part except in very few titles. I omitted 6900K due to its absurd price tag( immediately not an option for comparison) and i5 6/7 series due to the fact they will go up against 8T Zen for the roughly same price. So any way you look at it AMD is in a really good position. And then there is the dark horse SKU range, the best bang for the money, 6C/12T parts, if we are lucky enough - unlocked. These will just bury 6700K/7700K.

Your words sounds familiar...lol...hmm let me think when ive heard them before...oh yeah...bulldozer release lmao...yeah but when software will use more cores bulldozer will stomp everything...that was 5 years ago...and here we are...almost 2017 and an i3 humiliates bulldozer...

Same story repeats himself over an over again...
I don't even know what to say to this. Are you really comparing Bulldozer to Zen? After all we have seen thus far? Wow.
 

zinfamous

No Lifer
Jul 12, 2006
110,512
29,098
146
Most definitely. I'll still give it a few grains of salt since it's still more or less a rumour (albeit a more substantiated one).

I should've elaborated - I was referring more to his conclusion that benches from beta/ES chips are fake. Obviously they're not necessarily indicative of the final product, and I'm assuming they tested what they said they tested, but given the source I'd still put more faith in their results than, say, extrapolating performance from Blender run comparisons (although doing that has its own merits).

The only outright fake bench that has surfaced so far has been the one posted in the "AMD 'benchmark'" thread--renamed to post those fakes, oddly enough--from that Chinese forum, and one of those proven fake benchmarks remains at the top of the edited OP as if it was somehow legit. Go figure. :D
 

lolfail9001

Golden Member
Sep 9, 2016
1,056
353
96
The problem will be that you will have a choice between similarly priced 8C/16T Zen clocking to ~4-4.3Ghz Vs 4C/8T 7700K clocking to 4.6-4.8Ghz.
You look overly confident on pricing, aren't you?
So any way you look at it AMD is in a really good position. And then there is the dark horse SKU range, the best bang for the money, 6C/12T parts, if we are lucky enough - unlocked. These will just bury 6700K/7700K.
Considering the whole CCX thing, it is unlikely 6c/12t will ever come to be that year, to begin with.
 

ultima_trev

Member
Nov 4, 2015
148
66
66
OTOH, it performs nearly 10% *worse* in games than the hex core HEDT chip, 30% worse than a 6700k, and no better than an i5. If one accepts the data from this review, I think the power consumption looks excellent, the productivity benchmarks are pretty good (in line with what I expected), performance per watt in productivity tasks is very good, while I dont really understand how anyone can look at the gaming benchmarks and think they are good news. Granted, the clockspeeds are low, but the hex core intel has at least 20% more clockspeed headroom as well.

The 6800k also has an 8% higher turbo.

AMD giving us performance and power consumption between Haswell and Broadwell is more than most thought possible and they will likely undercut Intel on prices too. This will be great for the consumer.
 

Sweepr

Diamond Member
May 12, 2006
5,148
1,142
131
The only outright fake bench that has surfaced so far has been the one posted in the "AMD 'benchmark'" thread--renamed to post those fakes, oddly enough--from that Chinese forum, and one of those proven fake benchmarks remains at the top of the edited OP as if it was somehow legit. Go figure. :D

Right, I'm starting to think that CB score might be too optimistic and not reflective of overall performance, compared to Intel's quads at least. According to the latest benchmarks 8C/16T Ryzen at 3.2-3.3 GHz all core Turbo would edge stock 7700K with half the cores by only 10-12.5% in applications and still lose badly in CPU limited games - and that's before you overclock the latter to 5 GHz (easily done in many reviews). Go figure. :p

lolfail9001 said:
You look overly confident on pricing, aren't you?

That's a lot of 'assumptions' considering that user has zero info about Ryzen prices, overclocking headroom and only a limited idea of how it performs. I'm particularly curious to see how the 6C/12T Ryzen 'will bury 7700K' prediction holds up.
 
Last edited:

beginner99

Diamond Member
Jun 2, 2009
5,208
1,580
136

rainy

Senior member
Jul 17, 2013
505
424
136
Right, I'm starting to think that CB score might be too optimistic and not reflective of overall performance, compared to Intel's quads at least. According to the latest benchmarks 8C/16T Ryzen at 3.2-3.3 GHz all core Turbo would barely edge stock 7700K with half the cores by 10-12.5% in applications and still lose badly in CPU limited games - and that's before you overclock the latter to 5 GHz (easily done in many reviews). Go figure. :p
.

What a "lovely" another trolling attempt from your side.
7700K clocks are 4.2-4.5 GHz, when tested sample of Zen was run at 3.15-3.2/3.3 GHz - it's obvious that in programs which do not benefit from higher number of cores, Kaby Lake would be faster because of higher IPC and significant clock advantage.
 

bjt2

Senior member
Sep 11, 2016
784
180
86
Oh I have no doubt that intel parts will have the max OCing edge over Zen. The problem will be that you will have a choice between similarly priced 8C/16T Zen clocking to ~4-4.3Ghz Vs 4C/8T 7700K clocking to 4.6-4.8Ghz. Or even worse, 2x cheaper 4C/8T Zen clocking to 4-4.3Ghz VS 7700K. In games the 4C/8T highly clocked Zen will likely perform better than 8C part except in very few titles. I omitted 6900K due to its absurd price tag( immediately not an option for comparison) and i5 6/7 series due to the fact they will go up against 8T Zen for the roughly same price. So any way you look at it AMD is in a really good position. And then there is the dark horse SKU range, the best bang for the money, 6C/12T parts, if we are lucky enough - unlocked. These will just bury 6700K/7700K.


I don't even know what to say to this. Are you really comparing Bulldozer to Zen? After all we have seen thus far? Wow.

Let's do a little bit math:

6900 8c 3.2/3.7 140W -> 7700 4c+GPU 4.2/4.5 91W
Zen 8c 3.4+/??? 95W -> Zen 4c WITHOUT GPU 95W, how much? 4Ghz? Why? Why not at least 4.2/4.5 as INTEL? Because low power process? And the 8c that has higher clock and lower power? It's magic?