http://ark.intel.com/products/93790/Intel-Xeon-Processor-E7-8890-v4-60M-Cache-2_20-GHz
https://newsroom.intel.com/press-kits/intel-xeon-processor-e7-v4-family/
>24 cores
THANK YOU BASED INTEL
>>54943172
can it run moorhuhn?
>>54943172
MOAR COARS
>>54943208
>moorhuhn
goddamn how did you remember that shit?
>>54943172
Can't wait to put 4 of these in my DL580 G9.
>>54943285
What are you doing with that server?
>>54943356
If it were mine I do whatever I damn well please and could.
>>54943172
>24 cores
>48 threads
Put 4 of these in a server and you get 192 threads. Is there anything that can even scale on so many threads?
>>54944082
3D rendering?
>>54944082
>Is there anything that can even scale on so many threads?
Software developed for servers.
Wouldn't mind having just one of these, should make x265 encoding go pretty fast, practically real-time. :D
Yes of course it'll be faster than real-time, that was sarcasm you boneheads.
Noice
https://browser.primatelabs.com/geekbench3/5970915
>>54943172
Nice! Going to buy it right away!
Quick someone post the hyberbibeline comic
>>54944082
>Put 4 of these in a server and you get 192 threads.
installing gentoo on that must be a really big turn-on
>>54944162
>see results
>sheeeeeeyyyyyyiiiiiiiiiiiiiittttttttttt
Wow.
>>54943172
Could someone explain the Xeon branding to me? What is the difference between an e3, e5, and e7? Is it the same good/better/best branding as the i3/i5/i7 line?
>>54944197
Semi-relevant:
https://www.youtube.com/watch?v=v8hKjPh5SVo
>>54944162
>geekbench
Is this even vaguely meaningful? I know all synthetic benchmarks suck, but is geekbench one of the better ones?
>>54944107
>Yes of course it'll be faster than real-time
It won't be unless you do an extremely shitty job.
>>54944351
On closer inspection, there has GOT to be something wrong with this benchmark. See here:
https://browser.primatelabs.com/geekbench3/6294391
A 6-core Mac got over FOUR TIMES the score of this new CPU. And the Mac was, for some bizarre reason, running a 32-bit benchmark.
Something's fucky.
>>54944399
Pretty sure that result is fake
>>54944419
Then how can I trust anything else on the site? That result is off-the-walls bonkers, but they didn't catch it.
And that's only one result. Look through all the top results and you see similarly absurd scores.
https://browser.primatelabs.com/geekbench3/multicore
Of note:
>android phone crushes $40000 servers
https://browser.primatelabs.com/geekbench3/6659496
>144-core server barely beats the CPU OP posted at multithreaded performance
https://browser.primatelabs.com/geekbench3/5471006
There is something very badly wrong with this picture.
>>54943172
>24 cores
>48 fucking threads
>60mb L2
>3tb max memory
Just holy shit. I wouldn't know what to fucking do with it but I want two of these.
Only $7000 freedom dollarydoos.
>>54944399
The Xeon result is accurate, the others are faked/hacked/bloated/basically cheats to get to the top of the leader boards.
People are fucking stupid, it's really that simple.
>stupid fucking people will be the death of us all
>>54944624
Just wait for Linus to be sent 4 of them to put in a gaymen machine
>>54944275
More or less. Just mirrors the "i" model somewhat, but an e3 is more like an i5 or i7, and nothing consumer wise matches e5 and e7.
>>54944763
So how can I tell which are fake and which are real?
How is it that this one CPU competes with 64+ core servers? Either this CPU is years ahead of everything else or the benchmark is broken.
>>54944793
kek I made a similar comment about Linus in a thread about the newer kingston SSDs.
>>54944824
The Xeon score is for 4 PROCESSORS in the machine, 4 individual Xeons, so divide it by 4 and you like ~38K for just one of them which is pretty fucking accurate overall for something with 24 cores/48 threads based on comparing that with other processors like the last generation 12 core24 thread Xeon.
That is the single most powerful processor available as of today, that Xeon with 24 cores/48 threads so if you see anything past roughly ~40K for multi-core performance from a single physical CPU it's bullshit.
There's nothing on the market (or the planet for that matter) right now today in a single physical package as powerful as that Xeon E7-8890 v4.
>>54944811
E3 is unprocessed, e5 is dual processor and e7 is quad or octi processor capable. Depends on user requirements which one is ideal for your use basis. But yes, that's basically what it boils down to e3/e5/e7 : i3/i5/i7 enterprise : consumer
>>54944275
>>54944811
>>54944926
For what it's worth:
eX = Enterprise
iX = Individual
The E3s are non scalable, only one per machine.
>>54944275
Xeons are i7 processors with two changes:
1) The Xeon version of the processor will have a lot more L2/L3 cache, typically these days it's got massive amounts of L3 vs the same processor engineered as an i7 for consumer use
2) Xeons are SMP capable which means you can have more than one of them in a machine and that means more than one physical processor. For example, that SuperMicro Geekbench result posted above was done with a motherboard/system that has 4 Xeon 24 core/48 thread processors in it running (so 96 cores because 4 x 24 = 96 and 192 threads because 4 x 48 = 192).
Fucking beast machine unlike many others.
But just think, in 5-8 years we'll have single processors that will be that powerful, go figure.
>>54945120
I meant uniprocessor, it got auto corrected. I never knew what the actual meaning for the i was. Makes perfect sense now.
>>54945140
>But just think, in 5-8 years we'll have single processors that will be that powerful, go figure.
Not sure about this anymore. The server room really needs stuff like pci to get worked on now. That plus less heat and better energy efficiency for mobile makes me think raw power gains will take a back seat for a bit.
>>54944082
sap hana?
computing your mother's mass?
>>54945156
It'll happen. Intel has a working 96 core 192 thread prototype testing today (a Xeon, of course, with under 115W TDP), heard about it this morning. They're not resting on laurels, not in the least, and advancing the tech considerably.
>>54945156
>raw power gains will take a back seat for a bit.
It already has, judging by the minimal gains over the past few years. It's too easy to just slap more cores in right now plus AMD is offering little competition.
>>54944082
It's an 8xxx-series chip, so you can actually have them 8-way, or 192 cores and 384 threads.
>>54944891
>most powerful processor
>x86
>>54943172
Cucks.
>>54945213
What's better than that beast?
>>54945234
Is that a Sandy rig?
>>54945234
>all that unused ram
>all those unused cores
why?
>>54945696
sparc m8?
also
http://www.anandtech.com/show/9193/the-xeon-e78800-v3-review
>that autistic sperg fight in the comment section
kek
>>54946114
Are they niche hardware nowadays? I thought x86 dominates the server market simply because of the abundance of hardware and whatnot.
>>54944624
>>54943172
It is 5x cheaper to use older generation xeons http://www.techspot.com/review/1155-affordable-dual-xeon-pc/
>>54947164
don't let the plebs know, I want to still get that shit for cheap
>>54947250
>still
good luck finding a dual 2011 mb without getting jewed
>>54943172
>no turbo boost 3.0
Go into to trash!
kek
>>54944162
>Linux 2.6.32-573
Why these people is so estupid? Run a recent hardware with outdated shit kernel.
>>54943172
>18 of those cores were fired for pajeet cores
Ill pass thank you.
>>54947164
I know that but businesses don't tend to buy second hand shit because of warranty and on the spot service.
![s-l1600[4].jpg](https://i.imgur.com/vgB3fgS.jpg "s-l1600[4].jpg")
yoooo
These just came onto ebay today
http://www.ebay.com/itm/Dell-Precision-T5600-2x-2-60GHz-8C-E5-2670-32GB-RAM-2x-1TB-HD-Win7Pro-/182157878012?hash=item2a6974a2fc:g:H4YAAOSwQupXVZpG
I am severely, SEVERELY tempted to get one. I already own an HP z800 with 2x X5650 but god damn this amount of power for this cheap is calling to me.
This system is basically the same as the PC built in this post>>54947164
as well as that popular youtube video about the e5-2670.
>>54945725
>tolerance is bad
