• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Intel Broadwell Thread

Page 140 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Xeon E5 v4 'Broadwell-EP' Processor Family Officially Launched - Part 2

intel-xeon-e5-v4-sku-chart.jpg


intel-xeon-e5-v4-table.jpg


intel-xeon-e5-v4-vs-v3-performance-2.jpg


www.nextplatform.com/2016/03/31/broadwell-brings-xeon-e5s-balanced-performance-bump
 
AnandTech's review is out:

AnandTech - The Intel Xeon E5 v4 Review: Testing Broadwell-EP With Demanding Server Workloads

The fact that only 22 of those 24 cores are activated in the top Xeon E5 SKU is purely a product differentiation decision. The 18 core Xeon E5 v3 used exactly the same die as the Xeon E7, and this has not changed in the new "Broadwell" generation.

The largest die (+/- 454 mm²), highest core (HCC) count SKUs still work with a two ring configuration connected by two bridges. The rings move data in opposite directions (clockwise/counter-clockwise) in order to reduce latency by allowing data to take the shortest path to the destination.

...Just like Haswell-EP, the Broadwell-EP Xeon E5 has three different die configurations. The second configuration supports 12 to 15 cores and is a smaller version (306mm²) of the third die configuration that we described above. These dies still have two memory controllers.
Click to expand...

24-core die size: 454mm²
15-core die size: 306mm²

On Haswell, one AVX instruction on one core forced all cores on the same socket to slow down their clockspeed by around 2 to 4 speed bins (-200,-400 MHz) for at least 1 ms, as AVX has a higher power requirement that reduces how much a CPU can turbo. On Broadwell, only the cores that run AVX code will be reducing their clockspeed, allowing the other cores to run at higher speeds.
Click to expand...

v4_AVXandTurbo_575px.png


67413c.png


67413k.png


www.anandtech.com/print/10158/the-intel-xeon-e5-v4-review
 
I'm actually kind of shocked that a single 22C Broadwell-EP is only slightly behind a dual Ivy Bridge-EP set-up with 24 physical cores and much higher clockspeeds.
 
BTW the die size for the 10-core parts that should power Broadwell-E (Core i7-69xx) is ~246mm². 22nm Haswell-E packs 8 cores at 356mm².

Here's Cinebench R15 with a 2S Xeon E5-2699 v4 system:

14-1080.2637359778.jpg


And benchmark results from ServerTheHome:

Intel-Xeon-E5-2699-V4-c-ray-Benchmark.png


Intel-Xeon-E5-2699-V4-7-zip-compression.png


Intel-Xeon-E5-2699-V4-NAMD-Benchmark-800x499.png


www.servethehome.com/intel-xeon-e5-2699-v4-benchmarking-the-top-end

If you missed last page:

Sweepr said:
Xeon E5 v4 'Broadwell-EP' Processor Family Officially Launched

xeon%20ep_zps65cji2ti.jpg


xeon%20ep2_zpsm2m4tsfb.jpg


Three die configurations: 10-core (up to 8 active), 15-core (up to 14 active) and 24-core (up to 22 active)

kl%20die_zpswntfmqoh.jpg


More here

Supermicro New Generation Servers with Intel Xeon E5 2600 V4 (Broadwell) Product Family Support

SPECjbb2015 Results for Xeon E5-2699 v4

Amazon listing

MSI's X99S XPOWER AC LGA2011-v3 motherboard gets official support after BIOS update

intel-xeon-e5-v4-sku-chart.jpg


intel-xeon-e5-v4-table.jpg


intel-xeon-e5-v4-vs-v3-performance-2.jpg


www.nextplatform.com/2016/03/31/broadwell-brings-xeon-e5s-balanced-performance-bump
Click to expand...
 
Arachnotronic said:
You can just feel the disdain for Intel's server product superiority ooze from this reviewer.
Click to expand...
Not to mention those regular senteces ending with".."

I don't think Intel has many problems. I see from what I can tell a pretty nice set of improvements / new features, more cores, new production process. Intel always said they promise 20% per gen, which they pretty much deliver.

And never mind Skylake that's coming next year..
 
JDG1980 said:
One die is smaller than Tahiti or Tonga, the other is no bigger than Hawaii. That demonstrates just how badly we're getting soaked on pricing, due to the lack of competition.
Click to expand...

Cost/mm^2 here is much higher, your comparison doesn't make a ton of sense. Also can guarantee you that the R&D intensity of developing a high performance server processor is much greater than that of developing a dGPU.
 
JDG1980 said:
One die is smaller than Tahiti or Tonga, the other is no bigger than Hawaii. That demonstrates just how badly we're getting soaked on pricing, due to the lack of competition.
Click to expand...

Yes and NAND is even cheaper. OMG Evil Nvidia and AMD!

Just shows how little you know about cost and business. Specially R&D.
 
I wonder if we will see a full 15-core SKU based on the MCC die with Xeon E5-4600 v4 (or maybe even Xeon E7 v4).
With a 14 core SKU, Intel can turn off the leakiest or slowest core of the die.
 
Last edited:
witeken said:
Not to mention those regular senteces ending with".."

I don't think Intel has many problems. I see from what I can tell a pretty nice set of improvements / new features, more cores, new production process. Intel always said they promise 20% per gen, which they pretty much deliver.

And never mind Skylake that's coming next year..
Click to expand...

Intel is claiming that too. But I dont see it. The ipc gain is only about 5% and little or no gain in clockspeed. The only way they get significant gains is with more cores.

Makes me wonder if BW-E will improve at all over HW-E.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top