Processador AMD ZEN2 7nm (ler primeiro post)

Deveria ser gralha do Anandtech, porque todos os rumores apontavam para o 3500X e o 3500 não terem SMT. Fui ver agora vários sites e está tudo sem SMT. Por exemplo o Toms:
LVeEyfR.png


Meanwhile, the Ryzen 5 3500X slots in as a six-core model but comes without simultaneous multi-threading (SMT), meaning it only has six threads of performance on offer. AMD tends to avoid de-featuring its retail processors, so the lack of SMT stands out among the Ryzen 3000 product stack.
Clicar para expandir...

https://www.tomshardware.com/news/amd-release-ryzen-9-3900-ryzen-5-3500x,40595.html

Já agora, alegadamente, o 3500 é igual ao 3500X, mas em vez de ter 32 MB de L3, "só" tem 16 MB. Se tivesse que apostar, em aplicações Desktop, não se irá notar a maior parte das vezes......
Vamos ver se o 3500 não será apenas "OEM Only", para não ter que ser preciso conhecer alguém na Máfia Italiana para se arranjar um. :D
 
Estou agora com um R5 2600 e em benchs o 3600 arrebenta mesmo e estou com os testes com a board anterior, bios ABB e ram a 2933 :D

Mas curiosamente nos 3 jogos que testei, a diferença é muito pequena, no Shadow of TR até teve um pouco mais de média...
 
Fiz um exercicio simples. Somei as frames todas do 3500x e do 9400f para ver onde estava essa vantagem do 9400F, e realmente 1122.65 vs 1114.43, com menos 8.22 frames no total... O 9400F foi mais rápido em 4 dos 9 jogos, num deles 0,38 FPS a mais :). Perde em 2 jogos e empata em 3. Um dos que perde é o GTA V onde a intel costuma liderar, o que provavelmente indica um botleneck dos 6 cores...

O mais engraçado é no Blender ter uma performance em tudo superior ao 3600, será que alguém se esqueceu de actualizar uma flag (if CPU id = "ryzen 3500x" do result-X%). No single core do cinebench até o 3500x tem mais performance que o 9400f.
 
SideWalker disse:
6 vs 16 threads... Tirando o caso especifico do pessoal com monitores de 240Hz, não vejo grande dificuldade na escolha. :002:
Clicar para expandir...
Outra vez arroz???
Esta mais que demonstrando que grande parte dos jogos escalam mal com mais cores. Alias, o 9400f na maioria dos jogos fica á frente do 2700 e por vezes chega a 15%.
No caso especifico do wow num 9900k a passagem de 4c/4t para 8c/8t ganhou menos 10fps a 1440p com uma 2080.... de 139 para 147 fps. Em tudo o que se relacione com produtividade ... sim 2700 bate um 9400f. Agora dse falarmos do 3500x eu não tenho duvidas em escolhe amd.

LSeven, estavamos a falar do 2700.
 
PWP disse:
Neste momento compensa mais um R7 2700 ou um R5 3600? Ambos devem fazer 4.1 Ghz em all cores... O 3600 tem um aumento de IPC de quanto?
Clicar para expandir...

+/- 15%


Parece que o TRX40 não será compatível com os Threadrippers de 1ª e 2º Geração e os próximos Threadrippers não serão compatíveis com o anterior X399

AMD TRX40 Chipset Not Compatible with 1st and 2nd Gen Threadrippers

AMD is giving finishing touches to its 3rd generation Ryzen Threadripper HEDT processor lineup, and the first wave of these chips, starting with a 24-core model, will launch alongside the AMD TRX40 chipset. It turns out that the chipset won't be compatible with 1st and 2nd generation Ryzen Threadripper processors. The upcoming 3rd generation Threadripper chips won't be backwards-compatible with the AMD X399 chipset, either. We've been hearing from reliable sources rumors of this segmentation from AMD for a few days now, and tech journalist ReHWolution just tweeted its confirmation having obtained info on upcoming motherboards from ASRock. The company is working on the TRX40 Taichi and TRX40 Creator motherboards.

The underlying reason between this restriction remains a mystery. We know that the EPYC "Rome" MCM is pin-compatible with first-generation EPYC "Naples" chips due to the fact that the newer chips are drop-in compatible with older servers via a BIOS update. The TR4 socket, too, is nearly identical to SP3r2, but for four out of eight memory channels being blanked out. It remains to be seen if for TRX40 motherboards, AMD re-purposed these unused pins for something else, such as additional PCIe connectivity or more electrical pins. We'll find out in November, when AMD is expected to launch these chips.
Clicar para expandir...

https://www.techpowerup.com/259993/...compatible-with-1st-and-2nd-gen-threadrippers
https://twitter.com/ReHWolution/status/1181954537197854721

Interessante. Se percebo bem, o socket será o mesmo, por isso será interessante ver qual é a justificação. 8 canais de memória?
 
Não é exactamente uma novidade, mas do Robert Hallock:

AMD_Robert
Technical Marketing118 points · 3 days ago
No, it has a B550A motherboard. This is a version of the PCIe Gen 3 "Promontory-LP" (e.g. X470, B450) chipset specifically for use in pre-built systems (e.g. OEM customers).
Clicar para expandir...
https://www.reddit.com/r/Amd/comments/dghxdi/this_prebuilt_at_bestbuy_has_an_asrock_b550/

o B550A é específico para OEM, e um rebrand dos anteriores.

Já agora, Rome

UK Research and Innovation (UKRI) are pleased to announce that Cray have been awarded the contract to supply the hardware for the next national supercomputer, ARCHER2.

ARCHER2 should be capable on average of over eleven times the science throughput of ARCHER, based on benchmarks which use five of the most heavily used codes on the current service. As with all new systems, the relative speedups over ARCHER vary by benchmark. The ARCHER2 science throughput codes used for the benchmarking evaluation are estimated to reach 8.7x for CP2K, 9.5x for OpenSBLI, 11.3x for CASTEP, 12.9x for GROMACS, and 18.0x for HadGEM3. Needless to say, ARCHER2 represents a significant step forwards in capability for the UK science community, with the system expected to sit among the fastest fully general purpose (CPU only) systems when it comes into service in May 2020.

As has been previously announced, because ARCHER2 is being installed in the same room as the current ARCHER system, there will be a period of downtime where no service will be available. Therefore, users should plan their work appropriately. ARCHER is due to end operation on 18th February 2020, and ARCHER2 will be operational from 6th May 2020.
Clicar para expandir...

  • A peak performance estimated at ~ 28 PFLOP/s
  • System Design:
    • 5,848 compute nodes, each with dual AMD Rome 64 core CPUs at 2.2GHz, for 748,544 cores in total and 1.57 PBytes of total system memory
    • 23x Shasta Mountain direct liquid cooled cabinets
  • 14.5 PBytes of Lustre work storage in 4 file systems
  • 1.1 PByte all-flash Lustre BurstBuffer file system
  • 1+1 PByte home file system in Disaster Recovery configuration using NetApp FAS8200
  • Cray next-generation Slingshot 100Gbps network in a diameter-three dragonfly topology, consisting of 46 compute groups, 1 I/O group and 1 Service group
  • Shasta River racks for management and post processing
  • Test and Development System (TDS) platform, to be installed in advance
  • Collaboration platform with 4 x compute nodes attached to 16 x Next Generation AMD GPUs
  • Software stack:
    • Cray Programming Environment including optimizing compilers and libraries for the AMD Rome CPU
    • Cray Linux Environment optimized for the AMD CPU blade based on SLES 15
    • Shasta Software Stack
    • SLURM work load manager
    • CrayPat as profiler
    • GDB4HPC as debugger
Clicar para expandir...
https://www.archer.ac.uk/about-archer/news-events/archer2.php#191014
 
Inicie sessão ou registe-se para poder participar.
Back
Topo