See How Your Linux System Compares To A $300 Broadwell-EP CPU That Lacks Turbo Boost

[ossuser]

I mostly don't care these days. Speedwise even old Phenom X4 955 is quite adequate these days, if we ignore power consumption for a moment, and even there GPU is much more "burning" issue. At least on Linux, there is much more to gain from SW optimisation than HW upgrade, so I don't choose HW in order to jump over a particular obstacle, but to hit some broad computing power target that I deem needed and then tweak/rework the SW.

Zen benchmarks will be of interest to me since it represents qualitative jump ( new DDR4 with lower power consumption, higher freq / which is good for APU/ and higher density /16GiB-ones available, much lower power consumption on average, unified socket, new SSE and crypto engine etc etc.)

While still using a workstation board with Westmere Xeon (6-core 95W), I totally agree.

I'm also looking forward to Zen to see it will be a nice upgrade.

[sdack]

I mostly don't care these days. Speedwise even old Phenom X4 955 is quite adequate these days, if we ignore power consumption for a moment, ...

The point of Michael's article probably is exactly that - to see what this low-power, low-cost and multi-core CPU can do. It should be right up "your alley".

There is still need for some moderate CPU power on desktops when these don't get paired with modern GPUs. Most pressing issue these days for main stream consumers seems to be video playback. When the graphics hardware doesn't support H.265/HEVC video decoding and it needs to done in software (including decoding of 5+1 AAC audio channels), then the CPU load goes up. Older Pentiums and Celerons even with 2+ GHz cannot keep up. Although it's only a matter of time is it currently a new issue, because many web sites are dropping Adobe Flash video support for HTML5 video. While it is good to drop the Adobe "Trash" Player is the support for hardware-accelerated HTML5 video in Firefox for Linux still under development (for some it works, some can try to enable it manually and others go empty). In contrast to this have modern mobile phones no issues at all playing H.265/HEVC videos at 1080p.

[DanL]

This means that an "8 core" Zen is really more like two four-core chips that happen to be sitting on a piece of silicon and that have to communicate through the northbridge just like a 2 socket system would have to do.

With the advances in memory controllers and process tech, I'm not convinced that this is going to make a significant difference in performance. Zen is built "from the ground up" to minimize the disadvantages of this sort of design. The way you put it (especially using the word "northbridge") makes it sound like the old days of Intel dual cores where they hacked two Prescott or Cedar Mill cores together on a die to make make dual cores (Smithfield and Presler, respectively).

[sdack]

I thought that was what was done in most 8+ core designs too. NUMA? ...

Not all are like this. The first true quad-core was the AMD Phenom 9850, which unlike Intel's CPUs at the time wasn't just 2 duo-core dies glued together to make 4 cores. See here:

http://www.computershopper.com/var/ezwebin_site/storage/images/components/product-profile/amd-phenom-x4-9850-black-edition-processor/33572-1-eng-US/amd-phenom-x4-9850-black-edition-processor1.jpg

I found this was a beautiful design, although bugged at start and with a high power draw, was it based on a symmetrical design. The 4 cores are rotated and mirrored images of each other.

https://www.technic3d.com/article/pics/1466/Piledriver_Die.jpg

This is the die of the AMD Piledriver aka AMD FX 8350. Same here again with a symmetrical design only this time is it 4 modules with 8 cores (2 cores per module) where the modules are rotated and mirrored images of one another.

[Ericg]

Vs an i7-6700k

http://openbenchmarking.org/result/1608288-HA-1608277LO59

[chuckula]

Vs an i7-6700k

http://openbenchmarking.org/result/1...HA-1608277LO59

There's one test in there that's actually important for an 8 core server chip and that's the pgbenchmark. It looks like a winner from that perspective.

[Mark Rose]

Zen benchmarks will be of interest to me since it represents qualitative jump ( new DDR4 with lower power consumption, higher freq / which is good for APU/ and higher density /16GiB-ones available, much lower power consumption on average, unified socket, new SSE and crypto engine etc etc.)

Zen most likely won't be amazing at floating point throughput. Its 256-bit AVX2 runs at half speed. On the other hand, 256-bit AVX tends to be memory bandwidth bound on Intel's offerings, so the penalty might not be so great.

[DanL]

I mostly don't care these days. Speedwise even old Phenom X4 955 is quite adequate these days, if we ignore power consumption for a moment,

The point of Michael's article probably is exactly that - to see what this low-power, low-cost and multi-core CPU can do. It should be right up "your alley".

It seems like apples and oranges to me. The chip in the article isn't really aimed at desktop users. I don't really see why anyone but dedicated server users who could take advantage of LGA 2011 features (multi-CPU, large amounts of RAM) would want this CPU. For desktop, I think most users would be happier with a Skylake chip (core i5/i7 or Xeon E3v5).

I almost bought a Xeon E3 1260Lv5 and built a new system earlier this year, but the short supply of the low-power models and the crazy price gouging (upwards of $550 for a chip that lists $300) stopped me. I'll just keep chugging along with my gimpy Phenom II tri-core. About the only place I find myself wishing for more CPU speed is HEVC encoding, which I don't do a ton of.

[sdack]

I almost bought a Xeon E3 1260Lv5 and built a new system earlier this year, but the short supply of the low-power models and the crazy price gouging (upwards of $550 for a chip that lists $300) stopped me. I'll just keep chugging along with my gimpy Phenom II tri-core. About the only place I find myself wishing for more CPU speed is HEVC encoding, which I don't do a ton of.

H.265/HEVC encoding isn't done on CPUs any longer. The Nvidia Geforce 960 and the new Pascal 10x0-series are doing it in hardware and at such extreme speeds that CPUs have been overcome. If you got 10-20 fp/s from software encoding do you get 200+ fp/s from an Nvidia card (which can encode two videos in parallel, thereby doubling the actual possible speed when you have multiple videos to encode). It is such an extreme performance leap forward it is no longer funny trying to do this on a CPU in software.

[DanL]

If you got 10-20 fp/s from software encoding do you get 200+ fp/s from an Nvidia card. It is such an extreme performance leap forward it is no longer funny trying to do this on a CPU in software.

Wow. I didn't know it was such a dramatic difference. I have a GTX 950, so I'll have to play with NVENC when I get a chance (and the AC isn't on in the house). Unfortunately, my tool of choice has been Handbrake and the devs have no interest in NVENC.