• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD THREADRIPPER VS INTEL SKYLAKE X

N19h7m4r3 said:
I'm looking forward to Threadripper reviews; especially with PcPer doing core-core latency tests. I want to see how well two modules with 2 CCXs each copes with latency. Along with what memory frequencies the platform will hit on launch. I'd say launch with 3200Mhz Quad Channel would be fantastic for them if they can manage it; especially in gaming tests.
Click to expand...

Yeah also I'm interested to see if quad-channel memory helps the IF speed, or if it's solely RAM clockspeed related.

Hopefully with Zen2 they'll be able to make 3600 easy to achieve, and 4000+ with a good motherboard and overclocks.

Failing that, IF will essentially be perfected by DDR5 in 2020. As 5500+ MHz will be the 'normal' speed. And 7000-8000 MHz will be your premium overclocked speed.
 
Pricing for two Threadripper CPUs has been spotted on a press deck apparently.
Those are some nice boost clocks, I wonder what temps and overclocking will be like.

https://videocardz.com/newz/amd-ryzen-threadripper-1950x-to-cost-999-usd
LML85g16RGOj2v_lx8j-_Q.png
 
There are 3 SKUs below the 1920X (two 10c, one 12c) but there are also 4 SKUs between those two, spread over a $200 range. So we may still see a $600-650 entry-level part.
 
DragonQ said:
There are 3 SKUs below the 1920X (two 10c, one 12c) but there are also 4 SKUs between those two, spread over a $200 range. So we may still see a $600-650 entry-level part.
Click to expand...

I'm hoping for a decent 10-12 core with great base clocks and also 4-4.1Ghz Boost. Means less reason to bother overclocking.

Hoping the Noctua air coolers also launch same time so we can get rolling right away if these are good.
 
N19h7m4r3 said:
I'm hoping for a decent 10-12 core with great base clocks and also 4-4.1Ghz Boost. Means less reason to bother overclocking.

Hoping the Noctua air coolers also launch same time so we can get rolling right away if these are good.
Click to expand...

Well that'll be for 1-2 cores, there likely won't be a SKU any money which will clock above 3.6-3.7 All-core load. Need to overclock for that.
 
AllBodies said:
Well that'll be for 1-2 cores, there likely won't be a SKU any money which will clock above 3.6-3.7 All-core load. Need to overclock for that.
Click to expand...

3.6-3.7 base is plenty, then just overclock 4-6; possibly eight cores to 4Ghz and sorted for every workload, and some gaming.

All depends on temps; and of course how well the RAM in quad channel will work with overclocking, and affect performance.

I game at 1440p, so super high CPU clocks aren't essential, but in some MMOs and older games are still good to have.
 
Harlequin said:
https://www.bit-tech.net/news/tech/cpus/intel-badmouths-amds-epyc-design/1/

Intel are already on the attack, slating the way the cores work together , but neglecting to mention the `mesh` system their SLX uses , that requires the same optimisations
Click to expand...

Except from an optimisation point of view the Intel solution is worse/more limited.

With the Intel solution EVERY core talking to ANY core has high latency, and this latency isn't reduced by RAM speed (or not much anyway).

With the AMD solution talking within a 4-core CCX has VERY low latency (like Skylake-non-X level), and then only crossing CCX boundary has high latency (a bit above Skylake-X). And this latency is also reduced dramatically as RAM clockspeed increases.

So there's clearly more room for improvements and clever code solutions with AMDs approach. E.g. running many instances of a program, but limiting each instance to 4 cores each, this would then have about 1/3 the core-to-core latency of Intel's new chips.
 
AllBodies said:
Except from an optimisation point of view the Intel solution is worse/more limited.

With the Intel solution EVERY core talking to ANY core has high latency, and this latency isn't reduced by RAM speed (or not much anyway).

With the AMD solution talking within a 4-core CCX has VERY low latency (like Skylake-non-X level), and then only crossing CCX boundary has high latency (a bit above Skylake-X). And this latency is also reduced dramatically as RAM clockspeed increases.

So there's clearly more room for improvements and clever code solutions with AMDs approach. E.g. running many instances of a program, but limiting each instance to 4 cores each, this would then have about 1/3 the core-to-core latency of Intel's new chips.
Click to expand...

This ^^^^ AMD has clearly come up with the far better solution, given that Intel's 'Mesh' also doesn't address the low yield huge dies needed to make high core count CPU's.

Its interesting to see this behaviour from Intel, they are behaving like a spoilt child who has just been told for the first time in his life "NO"
There is an arrogance to the way Intel are conducting themselves with regards to AMD's challenge to them.

"How very dare you make a better product than ours, i'm going to tell everyone you smell funny"
 
humbug said:
This ^^^^ AMD has clearly come up with the far better solution, given that Intel's 'Mesh' also doesn't address the low yield huge dies needed to make high core count CPU's.

Its interesting to see this behaviour from Intel, they are behaving like a spoilt child who has just been told for the first time in his life "NO"
There is an arrogance to the way Intel are conducting themselves with regards to AMD's challenge to them.

"How very dare you make a better product than ours, i'm going to tell everyone you smell funny"
Click to expand...

As i pointed out yesterday, Intel just had their 24yr streak of being the biggest semiconductor manufacturer in the world ended by Samsung, to say they are feeling a bit butthurt is probably an understatement...

1) There reign at the top has come to and end bested by Samsung
2) They Skylake-X stuff is a joke and has not been received overly well
3) They are now mudslinging, name calling generally slandering AMD with dubious slides on their presentation for new products
4) They are about to release a confusing and overly complicated product stack in a bid to keep customers
5) They are probably massively upset that AMD and Samsung got together to work out a deal for Chips, as this seems to have done wonders for AMD's products (GPU's non withstanding)

All that time being the fat cat licking the cream and giving minimal effort is starting to fall down around them, other people have caught up, some have overtaken, so they seem to be in panic mode now and trying all sorts of tactics to keep their stuff together.
 
Didn't I read on here the CEO of Intel doesn't care much for desktop and retail parts now anyway? Hence why SLX is overvolted and overheated?
 
Threadripper specs are official now with TDP and confirmed pricing. I hope the RAM support is stellar on launch.

Some of our readers might argue that some of these answers have been exposed in leaks – from our perspective, getting confirmation from AMD (even on the minor points) is critical here. We’re going to have to wait until nearer the launch to get the answers to these questions. AMD have reaffirmed they are aiming for an early August launch, with more info at SIGGRAPH in a couple of weeks.
Click to expand...

http://www.anandtech.com/show/11636/amd-ryzen-threadripper-1920x-1950x-16-cores-4g-turbo-799-999-usd

AmSDT4GlQKqm-KgItO_fUA.png


II2f16lGTmuSrsMIaKCQIA.png



c4auK7YxRISZguomRGNK7A.png


LFnoxM3YRhak4fOx-BP5nA.png
 
Ingesting, with 20% more threads the $800 1920X scores 12% higher than the $1,000 7900X.

In last nights stream Ryan Shrout said "because Intel have better single core performance Intel's CPU would still be faster even in Multithreaded"

These guys are so Intel besotted... clueless.
 
You must log in or register to reply here.
Back
Top Bottom