EVGA GeForce RTX 3080 Ti FTW3 Ultra Review
F1 2021: PC graphics performance benchmark review
Asustor Drivestor 4 Pro (AS3304T) NAS Review
Colorful iGame GeForce RTX 3090 KUDAN review
Corsair Virtuoso RGB Wireless XT headset review
Galax HOF OC Lab Water Cooling 4400 MHz CL19 review
Scythe Mugen 5 Black air cooler review
KFA2 GeForce RTX 3080 Ti SG review
Mountain Everest Max keyboard review
Corsair iCUE 7000X RGB PC chassis review
ZEN3 Based AMD Epyc Milan Server Processors TDP up-to 280W
While Intel is still HUGE in the server market for processors, AMD the past two years have seen double-digit sales growth, as well, they offer faster, cheaper, and more secure products. That's said, the industry has been waiting on ZEN3 Epyc procs as these obviously also get improved IPC gains.
In all in one package with extensive DDR4 channel support and many PCIe Gen 4.0 lanes in the platform, it could mean a serious new chapter in the data-center industry. 3rd Gen EPYC Milan processors will use the same socket as Zen 2 EPYC Rome processors so companies looking to upgrade will be able to transition to the company's latest chips. EPYC Milan processors will offer up to 64 cores and 128 PCIe 4.0 lanes. To date, however, the specs of the Milan server procs have been hazy.
Specifications of the new AMD Epyc (Milan) processors now can be placed into an overview, the TDP of the new chips appears to stretch to 280W for the fastest clocked flagship model with 64 cores and 128 threads, as can be decrypted using the diagram below, a slide posted by ExecutableFix on Twitter. He also dispensed full specifications for the 3rd Gen EPYC series that now are nicely tabled up at Videocardz. AMD 3rd Gen EPYC Milan is now scheduled to launch in March. Below you can see the nomenclature and table with alleged processors and specs.
| AMD 3rd Gen EPYC Milan | |||||
|---|---|---|---|---|---|
| SKU | Cores/Threads | Base Clock | Boost Clock | L3 Cache | TDP |
| EPYC 7763 | 64C/128T | 2.45 GHz | 3.5 GHz | 256MB | 280W |
| EPYC 7713(P) | 64C/128T | 2.0 GHz | 3.675 GHz | 256MB | 225W |
| EPYC 7643 | 48C/96T | 2.3 GHz | 3.6 GHz | 256MB | 225W |
| EPYC 75F3 | 32C/64T | 2.95 GHz | 4.0 GHz | 256MB | 280W |
| EPYC 7543(P) | 32C/64T | 2.8 GHz | 3.7 GHz | 256MB | 225W |
| EPYC 7513 | 32C/64T | 2.6 GHz | 3.65 GHz | 128MB | 200W |
| EPYC 74F3 | 24C/48T | 3.2 GHz | 4.0 GHz | 256MB | 240W |
| EPYC 7443(P) | 24C/48T | 2.85 GHz | 4.0 GHz | 128MB | 200W |
| EPYC 7413 | 24C/48T | 2.65 GHz | 3.6 GHz | 128MB | 180W |
| EPYC 73F3 | 16C/32T | 3.5 GHz | 4.0 GHz | 256MB | 240W |
| EPYC 7343 | 16C/32T | 3.2 GHz | 3.9 GHz | 128MB | 190W |
| EPYC 7313(P) | 16C/32T | 3.0 GHz | 3.7 GHz | 128MB | 155W |
| EPYC 72F3 | 8C/16T | 3.7 GHz | 4.1 GHz | 256MB | 180W |
H83
Senior Member
Posts: 3490
Joined: 2009-09-08
Senior Member
Posts: 3490
Joined: 2009-09-08
#5882790 Posted on: 01/30/2021 04:07 PM
compared to server,intel is doing great in desktop.
they're getting pummelled in server/datacenter.even nvidia jumped to amd based systems for dgx and it must've been a hard pill to swallow.
Intel´s luck is that they are the only one with enough fab capacity to supply most of the demand, specially now...
compared to server,intel is doing great in desktop.
they're getting pummelled in server/datacenter.even nvidia jumped to amd based systems for dgx and it must've been a hard pill to swallow.
Intel´s luck is that they are the only one with enough fab capacity to supply most of the demand, specially now...
tunejunky
Senior Member
Posts: 1312
Joined: 2017-08-18
Senior Member
Posts: 1312
Joined: 2017-08-18
#5882812 Posted on: 01/30/2021 05:35 PM
and they fill the ecosystem with product
Intel´s luck is that they are the only one with enough fab capacity to supply most of the demand, specially now...
and they fill the ecosystem with product
tunejunky
Senior Member
Posts: 1312
Joined: 2017-08-18
Senior Member
Posts: 1312
Joined: 2017-08-18
#5882815 Posted on: 01/30/2021 05:48 PM
right as we "speak", AMD has standing orders from Google, AWS, M$, and most major players in "the cloud" services, Pharma, and Bio-Tech. Then there are the standing orders for the supercomputers using the Epyc Milan series and the major cable and internet providers in the USA, not to mention elsewhere like anywhere under the Paris Accords.
and they're less expensive, not just more powerful.
if anyone was still wondering about capacity at TSMC, please stop. they have more capacity at 7,5, and soon 3n than anyone else. other fabs have to output dies sequentially as they are the ones who have limited capacity. TSMC has (and is building more) enough fabs to output in parallel. and unlike other heavy industry industry changing the production line (for the same node) is easy by comparison.
once you hear about a shortage of Apple products or Android phones then you can complain about capacity.
right as we "speak", AMD has standing orders from Google, AWS, M$, and most major players in "the cloud" services, Pharma, and Bio-Tech. Then there are the standing orders for the supercomputers using the Epyc Milan series and the major cable and internet providers in the USA, not to mention elsewhere like anywhere under the Paris Accords.
and they're less expensive, not just more powerful.
if anyone was still wondering about capacity at TSMC, please stop. they have more capacity at 7,5, and soon 3n than anyone else. other fabs have to output dies sequentially as they are the ones who have limited capacity. TSMC has (and is building more) enough fabs to output in parallel. and unlike other heavy industry industry changing the production line (for the same node) is easy by comparison.
once you hear about a shortage of Apple products or Android phones then you can complain about capacity.
Kaarme
Senior Member
Posts: 2434
Joined: 2013-03-10
Senior Member
Posts: 2434
Joined: 2013-03-10
#5882833 Posted on: 01/30/2021 06:22 PM
if anyone was still wondering about capacity at TSMC, please stop. they have more capacity at 7,5, and soon 3n than anyone else. other fabs have to output dies sequentially as they are the ones who have limited capacity. TSMC has (and is building more) enough fabs to output in parallel. and unlike other heavy industry industry changing the production line (for the same node) is easy by comparison.
once you hear about a shortage of Apple products or Android phones then you can complain about capacity.
I wish someone told where the bottlenecks are so that I could blame the correct culprit! "Things just aren't working right now" or "demand is unexpextedly high" are too weird excuses. It's not like the PC business was born yesterday.
if anyone was still wondering about capacity at TSMC, please stop. they have more capacity at 7,5, and soon 3n than anyone else. other fabs have to output dies sequentially as they are the ones who have limited capacity. TSMC has (and is building more) enough fabs to output in parallel. and unlike other heavy industry industry changing the production line (for the same node) is easy by comparison.
once you hear about a shortage of Apple products or Android phones then you can complain about capacity.
I wish someone told where the bottlenecks are so that I could blame the correct culprit! "Things just aren't working right now" or "demand is unexpextedly high" are too weird excuses. It's not like the PC business was born yesterday.
Click here to post a comment for this news story on the message forum.
Senior Member
Posts: 2053
Joined: 2020-08-03
compared to server,intel is doing great in desktop.
they're getting pummelled in server/datacenter.even nvidia jumped to amd based systems for dgx and it must've been a hard pill to swallow.