Wednesday, March 24th 2021
Intel Readies Xeon W-1300 Socket LGA1200 Processors Based on "Rocket Lake"
Intel is reportedly giving final touches to the Xeon W-1300 line of enterprise processors targeting workstations, according to an ASRock CPU support list dug up by Komachi Ensaka. The processors are based on the same 14 nm "Rocket Lake" silicon as the company's 11th Gen Core desktop processors, and come in core counts of up to 8-core/16-thread. The lineup is expected to debut with five SKUs, three of which are 8-core/16-thread, and two 6-core/12-thread.
The lineup is led by the W-1370, with base frequency of 2.90 GHz, 16 MB of shared L3 cache, and 80 W TDP. Next up, is the slightly slower W-1390, clocked at 2.80 GHz, and 80 W TDP. The third 8-core part is the W-1390T, which is clocked at just 1.50 GHz (base), and comes with aggressive power-management that gives it a TDP rating of 35 W. The 6-core/12-thread W-1350P has the highest clock speeds, with a base frequency of 4.00 GHz, 12 MB of shared L3 cache, and 125 W TDP. The W-1350 is its slower sibling, clocked at 3.30 GHz, and 80 W TDP. The processors will be compatible with Intel Z490, W480, and H470 chipsets, besides their 500-series successors.
Sources:
ASRock, Komachi_Ensaka (Twitter)
The lineup is led by the W-1370, with base frequency of 2.90 GHz, 16 MB of shared L3 cache, and 80 W TDP. Next up, is the slightly slower W-1390, clocked at 2.80 GHz, and 80 W TDP. The third 8-core part is the W-1390T, which is clocked at just 1.50 GHz (base), and comes with aggressive power-management that gives it a TDP rating of 35 W. The 6-core/12-thread W-1350P has the highest clock speeds, with a base frequency of 4.00 GHz, 12 MB of shared L3 cache, and 125 W TDP. The W-1350 is its slower sibling, clocked at 3.30 GHz, and 80 W TDP. The processors will be compatible with Intel Z490, W480, and H470 chipsets, besides their 500-series successors.
22 Comments on Intel Readies Xeon W-1300 Socket LGA1200 Processors Based on "Rocket Lake"
ouch more, as Steve would put it, waste of sand
The only significant difference to W1200 is architecture, and avx512. so i guess if you need both ecc and avx512 prepare to cough up
Also price/performance is surprisingly good if you scale horizontally (multiple cheap servers) vs vertically (one expensive server), and you also have better redundancy This is an interesting one. Will it work the opposite way? Pair a server board with a desktop CPU?
Non of that means this is actually a good product though, (well unless you define good by salesfigures sure)
For example rog.asus.com/us/motherboards/rog-strix/rog-strix-x570-f-gaming-model/wtb
[INDENT=2]4 x DIMM, Max. 128GB, DDR4 4400(O.C)/4266(O.C.)/4133(O.C.)/4000(O.C.)/3866(O.C.)/3600(O.C.)/3400(O.C.)/3200(O.C.)/3000(O.C.)/2800(O.C.)/2666/2400/2133 MHz ECC and non-ECC, Un-buffered Memory *[/INDENT]
[INDENT=2]4 x DIMM, Max. 128GB, DDR4 3600(O.C.)/3400(O.C.)/3200(O.C.)/3000(O.C.)/2800(O.C.)/2666/2400/2133 MHz ECC and non-ECC, Un-buffered Memory *[/INDENT]
[INDENT=2]4 x DIMM, Max. 128GB, DDR4 3200(O.C.)/3000(O.C.)/2800(O.C.)/2666/2400/2133 MHz ECC and non-ECC, Un-buffered Memory[/INDENT]
[INDENT=2]* Refer to www.asus.com for the Memory QVL (Qualified Vendors Lists).[/INDENT]
[INDENT=2][/INDENT]
[INDENT=2]3rd Gen AMD Ryzen™ Processors[/INDENT]
[INDENT=2]2nd Gen AMD Ryzen™ Processors[/INDENT]
[INDENT=2]4 x DIMM, Max. 128GB, DDR4 MHz Un-buffered Memory *[/INDENT]
[INDENT=2][/INDENT]
[INDENT=2]2nd and 1st Gen AMD Ryzen™ with Radeon™ Vega Graphics Processors[/INDENT]
[INDENT=2]4 x DIMM, Max. 128GB, DDR4 MHz Un-buffered Memory[/INDENT]
[INDENT=2][/INDENT]
[INDENT=2]Dual Channel Memory Architecture[/INDENT]
On ECC memory support, my ROG X570 Strix beats my old ASUS ROG Strix X299 (for Intel Skylake X/CascadeLake X).
Come on, Intel - We need you to not suck in 2021. TSMC's over capacity already.
You forgetting the fact that previous gen Epyc still exist, or Threadripper in this regards ( workstation class ), heck even previous desktop class CPU would do just fine.
I think Intel targeting entry level, small SOHO with this CPU. With that in mind ( compatible with B or H chipset ) you can say goodbye to ECC/ IPMI. To be honest, 8 core will not sufficient to load security module and hypervisor. Much I care about security, I doubt that many administrator at this level would take advantages of IPMI or BMC ( AMD's IPMI ).
Seriously, it's a significant step up in performance per core, which matters a lot to most workloads. And don't forget workstations. Dell, HP, Lenovo etc. sells a lot of these. That's totally up to the motherboard maker.
E.g. a W480 motherboard like Supermicro X12SAE supports Pentiums, Celerons etc. The same goes for their other W480 boards. Only in theory.
Just enabling ECC is pointless, the memory controller needs to support it and be validated, otherwise it's just placebo.
I would love if AMD made CPUs like 5800W with ECC support and validated for 24-7 load. I'll happily pay $50 extra for that. Time to go to the eye doctor then ;)
Does all that make sense? ECC is a somewhat complicated thing.