Intel Broadwell-E Core i7 chip to feature 10 cores, 20 threads

[username2]

That graph assumes only one single kernel of one single program running across all cores.

My work PC, a bit busy:

top - 15:54:42 up 35 days, 6:03, 18 users, load average: 10.98, 5.05, 2.00
Tasks: 402 total, 20 running, 382 sleeping, 0 stopped, 0 zombie

Yeah, I could use 10 cores. I max out 6 cores/12 threads on a daily basis.

You said you did max out at your work, so probably in the IT industry it would be great but for home users does it make a big difference?

[mapesdhs]

DemonHighwayman writes:
> I picked up Crysis a couple weeks ago in a steam sale ...

I just bought some original packs on eBay, only about 2 or 3 UKP each, good to have
the proper booklet, etc. I include them with PCs I build for people.

Still playing it myself, and Warhead.


> and only just get over 65fps average at 1440p using a gtx 970 and 4790k. ...

I was using two 580 3GB SLI with very modified settings (extreme draw distance, better
shadows, etc.), gave about 45fps @ 1920x1200 (running with a 5GHz 2700K). Upgraded to a
single 980, went up a decent amount. However...


> When Crysis first came out I could run it at a decent frame rate (can't remember what
> but it was smooth) at 1680x1050 on my 8800 Ultra and Q6600. So hilariously enough it's
> probably a fair question with todays higher resolutions.

... thing about Crysis is the engine was not really that good. For ages after it came out,
everyone just assumed because it punished GPUs so hard that somehow that meant it was an
inherantly useful benchmark, but Warhead proved the original game's engine was not so well
optimised (Warhead runs noticeably and measurably better).

I also get some very odd numbers when trying to test configs with the Crysis Benchmark Tool,
as if it's ignoring some of the settings. I started doing various tests, but stopped after
some initial runs as some configs just didn't make sense (there's a 7970 CF result here
which shows what I mean, but I was seeing the same thing with SLI setups):

http://www.sgidepot.co.uk/misc/crysis.txt


Anyway, I plan on moving up to 2560x1440 at some point, but for that I'll switch to a 980 Ti.


As for BW-E, it boils down to pricing. Intel made the 5960X too expensive. Given the 3930K
was an 8-core CPU with 2 cores disabled, we already know they can sell these CPUs much
cheaper than they are. If Intel wants to revive the PC desktop platform, then top-end CPU
pricing needs to come down to sensible levels, though that seems unlikely given the crazy
price they set for the 6700K (who buys that thing in place of a used SB-E/IB-E or new HW-E?).

Ian.

[mapesdhs]

(null; is it possible to delete posts here?)

[DanceswithUnix]

You said you did max out at your work, so probably in the IT industry it would be great but for home users does it make a big difference?

Depends on the home user. For most 4 threads is plenty but for anyone running home VM farm experiments, large compile jobs, video encoding, bulk encryption or any heavily threaded workload then 20 threads isn't much.

[WritersBlock]

Damn, a reason to upgrade

[Mchotpoon]

Will this mean more cores heading to I5s...

[kalniel]

Will this mean more cores heading to I5s...

No. There are no core I5 Broadwell-Es.

[ALIAKSANDR]

Nice, but expensive ...

[mud_z]

It would be better if power would be like 14W.. I am waiting Intel.. Waiting...
you can do it INTEL. I believe in 2-3 years I can see that happening..

[qasdfdsaq]

I'll hang on for Skylake-E - hopefully by then we'll see 12cores+ so that I'll actually get a doubling from my Westmere, otherwise I can't see the point in upgrading. But finally we're moving on from 8...

6 cores will already give you double the performance of a Westmere

[qasdfdsaq]

I remember back when a flagship product launched about the same time as the mainstream, not over a year later and it's getting longer.
I think Nehelem was the last in fact. Since then intel have been tech that's a generation old with the same naming nomenclature as the current mainstream ie SB-EX was 2xxx gen tech but released as 3xxx.

The -E and -EP/EN platforms are bigger, more advanced and more complex chips so it's absolutely natural for them to take more development time. the -EX platform is the same again, with even more advanced features, and requires a further year of testing and development. The tech is most certainly not a generation old. It's the absolute latest there is in that platform. And that's a high end enterprise platform that requires considerable testing and validation.

Would you rather they pointlessly delayed the mainstream product for two years then, just to satisfy your pickyness? I wouldn't. (Given the ignorance reeking from your post it's unsurprising you've completely overlooked the EX series which comes out a year after the E series).

[GoNz0]

PCI lanes info would have been nice :-(

was going to ask the same.

Needs more PCIe lanes. We had 40 with 6-cores back with SB-E. With the rise of M.2, etc., 40 isn't enough anymore.

The 8 core 5960 still had 40 lanes so I doubt the new 8 core will get any more, maybe the 10 core.

It would be better if power would be like 14W.. I am waiting Intel.. Waiting...
you can do it INTEL. I believe in 2-3 years I can see that happening..

Would be nice if this was aimed at the server market and/or the HTPC crowd but it is quite the opposite.

[crossy]

Depends on the home user. For most 4 threads is plenty but for anyone running home VM farm experiments, large compile jobs, video encoding, bulk encryption or any heavily threaded workload then 20 threads isn't much.

20 threads not much on a home VM farm - what the heck are they doing? I can run a Windows VM and three Linux VM's quite happily on a rinky-dink C2D and that's what ... 4 threads? Similarly in a non-shared setup wouldn't you hit storage/memory limitations on compiles before you used up your 20 threads? Not posing this as criticism merely asking the question, especially as I'm planning to up my sights from running a mere 3VM's simultaneously.

So the US$64,000,000* question ... is it worth holding off on a Haswell-E purchase on the basis of this "rumour"? Or, is it more sensible to get a "cheap" (stop-gap?) 5820K while waiting for that more capable Broadwell-E drop-in replacement? Speaking of drop-in, kudos to Intel (if the rumour is true) for allowing an upgrade path for X99 owners, rather than inventing yet another socket. Asking the purchasing question because I'm thinking of splurging on a 5930K for a VM/media-coding setup this Christmas.

(* probably the list price of a 6950K knowing Intel!)

PS, is it just me or does the stated naming scheme not make a lot of sense? If I'd been in charge then instead of 6950, 6900, 6850 and 6800, I would have chosen 6900, 6880, 6865 (or 6870) and 6860.

[DanceswithUnix]

20 threads not much on a home VM farm - what the heck are they doing? I can run a Windows VM and three Linux VM's quite happily on a rinky-dink C2D and that's what ... 4 threads? Similarly in a non-shared setup wouldn't you hit storage/memory limitations on compiles before you used up your 20 threads? Not posing this as criticism merely asking the question, especially as I'm planning to up my sights from running a mere 3VM's simultaneously.

So the US$64,000,000* question ... is it worth holding off on a Haswell-E purchase on the basis of this "rumour"? Or, is it more sensible to get a "cheap" (stop-gap?) 5820K while waiting for that more capable Broadwell-E drop-in replacement? Speaking of drop-in, kudos to Intel (if the rumour is true) for allowing an upgrade path for X99 owners, rather than inventing yet another socket. Asking the purchasing question because I'm thinking of splurging on a 5930K for a VM/media-coding setup this Christmas.

(* probably the list price of a 6950K knowing Intel!)

PS, is it just me or does the stated naming scheme not make a lot of sense? If I'd been in charge then instead of 6950, 6900, 6850 and 6800, I would have chosen 6900, 6880, 6865 (or 6870) and 6860.

Compilers need megabytes per task, and that isn't much these days when 32GB of ram isn't crazy (even my home PC has 16GB). So, with enough ram to cache all your source code I find performance doesn't drop until I compile around 120 copies of gcc running at the same time so I think my single hard drive can cope with about a magnitude more threads than the Xeon at work can deliver. That is under Linux, last time I tried under Windows was in XP and I can only hope things have improved from that dire scaling.

Perhaps with the VMs you are right. I have seen people running big Java based systems on VMs which have a stupid number of threads going none of which seem to do much apart from consume ram and CPU. That probably isn't too common though.

[gingerninja7]

So, anybody want to lend me a grand to drop on one of these?

In all seriousness, the 6900K could be interesting - similar specs to the 5960X, but a ~10% higher base clock and better IPC, and potentially cheaper too (unless they add the 6950X at a new $1500 tier...).

As people have mentioned, 40 lanes looks less generous by the time you've added a PCIe drive to two...

[=assassin=]

It's nice that more cores are added at the top end, but it won't really affect my choices due to budget, which will be more aimed at the bottom of the pile. I'm guessing the lowest CPU, the 6800K, will be slightly higher in price than the 5820K initially.