- Genshin Impact (PC, PS4, Android, iOS)
- Diablo IV
- Előzetes élménybeszámolókon a Black Myth: Wukong
- World of Tanks - MMO
- Xbox Series X|S
- PlayStation 5
- Forza sorozat (Horizon/Motorsport)
- EA Sports WRC '23
- The Division 2 (PC, XO, PS4)
- Fortnite - Battle Royale & Save the World (PC, XO, PS4, Switch, Mobil)
Új hozzászólás Aktív témák
-
S_x96x_S
őstag
( AMD ThreadRipper Pro . Lenovo P620 ; 10 db PCie 4.0 SSD )
Achieving 11M IOPS & 66 GB/s IO on a Single ThreadRipper Workstation ( 2021-01-29 )
https://tanelpoder.com/posts/11m-iops-with-10-ssds-on-amd-threadripper-pro-workstation/Mottó: "A verseny jó!"
-
S_x96x_S
őstag
AMD 3rd Gen EPYC “Milan” full specifications confirmed ( launch in March )
https://videocardz.com/newz/amd-3rd-gen-epyc-milan-full-specifications-confirmed
8C/16T ... 64C/128T lesznek ...
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
az AMD vajon tényleg kiszervez pár dolgot ( APU, GPU) a Samsung 3nm-re? ( >= H2 2022 )
logika lehet benne, de ki tudja ..
1,5 évvel ezelőtti hir, hogy a Samsung licenszeli az AMD GPU technológiáját .. emiatt
mindenképpen át fogják ültetni a terveket, és legalább konzultációs szinten az AMD is be lesz vonva .. és ha a Samsung-os gyártási tapasztalatokat vissza portolják, akkor innen már csak egy kis lépés az AMD -nek is GPU-kat és APU-kat gyártani itt ...
Nagy kérdőjel volt az egész már annó :
"Something that’s equally ambiguous – but could have a big impact on how the deal works – is whether Samsung can use any of their existing IP with the new project, or for that matter who ends up owning any co-developed IP by the partnership. And there’s the big open question of whether AMD can use any IP improvements that were co-developed into their own products, or if they’ll be treated as a work-for-hire and stay with Samsung."a mostani pletyka:
"AMD could likely be one of the first companies to utilize Samsung's upcoming 3nm facilities."
AMD could outsource some of its APU and GPU production to Samsung in the near future
https://www.notebookcheck.net/AMD-could-outsource-some-of-its-APU-and-GPU-production-to-Samsung-in-the-near-future.517679.0.htmlMottó: "A verseny jó!"
-
S_x96x_S
őstag
Azért a rivális (jövőben várható) "Alder Lake-P" - proci egy kicsit "rémisztő" nekem.
nem is a teljesítmény, hanem a PCIe csatornák száma és típusa miatt.Alder Lake-P
CPU PCIe = 1x PCIe 5.0 x8 2x PCIe 4.0 x45.0-ból rögtön x8 -as ????
Kérdés, hogy lesz -e hozzá bele-passzoló nVidia kártya.persze a "Tiger Lake-H" - is tudja ugyanezt a sávszélt , csak még dupla annyi (x16) 4.0-ból
CPU PCIe = 1x PCIe 4.0 x16 1x PCIe 4.0 x 4Ha ezt a sávszélt a külső dokkoló is tudná, akkor reneszánszuk lenne a külső GPU/AI -s dokkolóknak.
Vajon az AMD mennyivel később jön ki a PCIe 5.0-val ???
Remélem csak túlaggódom és még az idén pozitívan meglepődök ...Mottó: "A verseny jó!"
-
S_x96x_S
őstag
drágább lesz az új EPYC .. ~ 6.2% -16.7% -al
Dell Canada Leaks AMD EPYC Milan Specs, Pricing
https://www.tomshardware.com/news/dell-accidentally-leaks-amd-epyc-milan-specs-pricing"""
The EPYC 7763 is one of many expected 64-core EPYC Zen 3 chips from AMD. This model in particular has a 2.45 GHz base clock, a 256MB L3 cache and a 280W TDP. In comparison to the existing EPYC 7H12, the EPYC 7763 will only be 6.2% more expensive, according to Dell Canada's pricing.
The EPYC 7713 shows identical specifications to the current EPYC 7662, but bear in mind that the first comes wielding Zen 3 cores. The EPYC 7713 could carry a price tag that's 16.7% higher than that of the EPYC 7662.
Overall, Milan doesn't appear to be more costly than Rome. From the processors that Dell has listed, we're looking at price escalation that spans between 6.2% to 16.7%. Considering the performance increases that should come along with the price changes, we're sure AMD won't have too much trouble finding customers.
"""Mottó: "A verseny jó!"
-
S_x96x_S
őstag
A Hetzner már a 16 magos "AMD's Ryzen™ 9 5950X" -t is felhősítette
( ~ havi bérléssel elérhető )
https://www.hetzner.com/dedicated-rootserver/ax101
persze ez annyira már nem újdonság mert
ugyanez a helyzet a többi AMD-s procival is
- AMD Ryzen 5 3600
- AMD Ryzen 7 3700X
- AMD Ryzen 9 3900"Mottó: "A verseny jó!"
-
S_x96x_S
őstag
-
S_x96x_S
őstag
https://www.asrockind.com/IMB-V2000
IMB-V2000 Mini-ITX Motherboard
AMD Ryzen™ Embedded V2000 SoC processors, up to 54W
-IMB-V2000P(V2748, 2.9GHz, 35-54W)
-IMB-V2000S(V2546, 3.0GHz, 35-54W)
-IMB-V2000M(V2718, 1.7GHz, 10-25W)
-IMB-V2000V(V2516, 2.1GHz, 10-25W)( már kb 1 hónapos hir .. ; de nekem újdonság )
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
A ZEN2 tudja, de a ZEN3 már nem ..
Ilyen dolgok miatt is fontos (lenne) ,
hogy minden fordító és kernel naprakész legyen a ZEN3-ból.ZEN2:
"The Zen 2 can mirror memory operands inside the CPU so that memory operands have no
latency at all in some cases."ZEN3:
"The advanced feature of mirroring memory operands inside the CPU that we saw in the Zen
2 (page 220) is no longer implemented in the Zen 3. This feature was probably very
expensive in terms of hardware complexity and temporary registers. The feature was mostly
useful in 32-bit mode where memory operands are more common. It is less valuable in 64-
bit mode where temporary variables and function parameters are mostly stored in registers.
Therefore, it makes sense to prioritize the hardware budget for other improvements instead."The microarchitecture of Intel, AMD and VIA CPUs ( By Agner Fog )
Last updated 2021-01-31.
https://www.agner.org/optimize/microarchitecture.pdfMottó: "A verseny jó!"
-
S_x96x_S
őstag
Mindfactory 2021 Január ..
( az 5900X , 5950X már nem is látszik a top20-ban .. )
https://imgur.com/a/T9McOFX#MHKthRd
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
Lehet, hogy mégis marad a TSMC -nél az AMD ?
Thursday 4 February 2021
AMD unlikely to change foundry partners for advanced nodes
AMD is unlikely to transfer its 5nm or 3nm chip orders from TSMC to Samsung Electronics, and its ongoing tight supply of processors is due mainly to ABF substrate shortages rather than insufficient capacity support from its current foundry partner,...
https://www.digitimes.com/news/a20210204PD201.htmlMottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz S_x96x_S #4747 üzenetére
> ABF substrate shortages
AMD Depends on TSMC, Who Depends on...TSMC manufactures the silicon itself and then packages it onto a substrate so the chip can be soldered or installed onto a motherboard of some type, though it isn't entirely clear which products TSMC packages for its customers or that use OSAT (outsourced assembly and test) firms for final assembly. This is called chip 'packaging,' and reports of chip packaging problems have been circulating for a few weeks.
Of course, it doesn't help that TSMC's capacity is already fully booked, but this news could shed new light on the reasons behind the supply issues. Previously, it was suspected that TSMC simply didn't have enough in-house lithographic capacity to make more chips, but it turns out that its reliance on external sources for either packaging supplies or final packaging services possibly contributes to the problem. TSMC's major ABF substrate suppliers, including Unimicron Technology, Kinsus Interconnect Technology, and Nan Ya PCB, are experiencing shortages.
Digitimes' sources indicate that AMD can currently only satisfy about 50 to 60 percent of Zen 2 demand for notebooks, and that ODM's (Original Design Manufacturers) expect to run into serious shortage issues when Zen 3-based notebooks enter the market in Q3. That's narrowed down to two possibilities (or perhaps both): Either TSMC or OSAT firms can't package enough chips, or AMD's chips for the PlayStation 5 are soaking up all of its production capacity.The situation isn't helped by the fact that demand is currently through the roof right now. AMD's CPUs, almost all GPUs, and the latest consoles are being bought up like pastries at a bakery — except there's only one bakery for an entire city.
Naturally, in such a situation, AMD is quite powerless as a fabless chipmaker because it relies on TSMC for its parts. One could argue that TSMC's suppliers should increase supply, but that would require increased capacity and investments from the substrate makers. No one can predict with any certainty that post-pandemic demand will remain at this level, potentially making that a poor investment for those firms, so that doesn't seem likely.
This means the current difficulties with finding Ryzen 5000, Radeon RX 6000, PS5, and Xbox Series X/S consoles are likely to continue. And the same goes for Nvidia's RTX 30-series GPUs.
https://www.tomshardware.com/news/amd-chip-shortage-packaging-issuesMottó: "A verseny jó!"
-
S_x96x_S
őstag
Ismét ...
Vajon mennyi esély van AMD-s ARM-es chipre?
( 5 éven belül szerintem nagy ! )December 9, 2020
Upcoming ARM Chip That's Faster Than Apple Silicon M1
https://singhkays.com/blog/arm-cpu-faster-than-apple-m1/
Amúgy ma 2 éves a Nuvia ...
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz Petykemano #4750 üzenetére
köszi, ez érdekes ..
> - zen3+: +4-7% IPC, AM5, DDR5
- ez vajon már PCIe 5.0 ?
- mennyi max cpu core-ra tervezték az AM5 foglalatot? ( 32 core? )Mottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz Petykemano #4753 üzenetére
> Kell azt előre megmondani, hogy mennyi lesz a magszám?
kíváncsiság, hogy mire terveznek ..
( 64 -lett ; maradhat ? )Mottó: "A verseny jó!"
-
S_x96x_S
őstag
The Latest Open-Source AMD Firmware / Coreboot Happenings In Early 2021
https://www.phoronix.com/scan.php?page=news_item&px=Open-Source-AMD-FW-2021
( FOSDEM - video , pdf )-----------------------------
"Most of the outside visible work by AMD around any open-source firmware / Coreboot support has been on the Chromebook front, thanks to seeing AMD powered Chromebooks and Google mandating the use of Coreboot. Thus as noted before there has been AMD Picasso work for Coreboot while still sticking to the binary AGESA and not much in the way of motherboard coverage outside of Chromebook devices.
Besides the Picasso work, there is support for AMD Cezanne and Majolica support under review for supporting the AMD Ryzen 5000 series mobile with Coreboot. Again, seemingly for upcoming Chromebook devices.
One remark made during the presentation was that, "despite there are no @amd.com patches going upstream, team contributing to Coreboot is officially hired by AMD" and "congratulations to our OSF friends who joined AMD, we already seethings moving faster."
Also exciting from 3MDEB, "There is a chance that AMD servers will also get [open-source firmware] support." As for that there is Project X as open-source Coreboot/Oreboot for AMD Zen. Prominent Coreboot developer Ron Minnich of Google has been working on that fully open-source Rust-based Oreboot support initially for AMD EPYC with the Rome reference board. But the hope is to ultimately support at least a more affordable developer board and see what happens but for now is a minimal implementation. Minnich's OSFC presentation from last year is embedded below for more details."Mottó: "A verseny jó!"
-
S_x96x_S
őstag
Lenovo ThinkCentre M75q-1 Tiny AMD Ryzen-based TMM Review
( Ryzen 5 Pro 3400GE )
https://www.servethehome.com/lenovo-thinkcentre-m75q-1-tiny-review-project-tinyminimicro/
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz Petykemano #4760 üzenetére
> A multithread áttörés még nem jött meg.
> Legalábbis a második, amikor már érdemes elgondolkodni 6-nál több magon.Azért >= RTX 3070 felett még mindig egy 8 vagy 12 magos a racionális ..
persze a tesztből kimaradt a 3080 és a 3060 ...
és nem lehet tudni, hogy az NVIDIA Resizable Bar - ja be lett -e kapcsolva ..
( valószínűleg nem )
https://www.techspot.com/review/2197-zen-3-cpu-gpu-scaling-benchmark/Mottó: "A verseny jó!"
-
S_x96x_S
őstag
friss ( persze ez még ZEN2-es )
"64 Cores of Rendering Madness: The AMD Threadripper Pro 3995WX Review"
https://www.anandtech.com/print/16478/64-cores-of-rendering-madness-the-amd-threadripper-pro-3995wx-review
"In the tests that matter, most noticeably the 3D rendering tests, we’re seeing a 3% speed-up on the Threadripper Pro compared to the regular Threadripper at the same memory frequency and sub-timings. The core frequencies were preferential on the 3990X, but the memory bandwidth of the 3995WX is obviously helping to a small degree, enough to pull ahead in our testing, along with the benefit of having access to 8x of the memory capacity as well as Pro features for proper enterprise-level administration.
The downside of this comparison is the cost: the SEP difference is +$1500, or another 50%, for the Threadripper Pro 3995WX over the regular Threadripper 3990X. With this price increase, you’re not really paying +50% for the performance difference (ECC memory also costs a good amount), but the feature set. Threadripper Pro is aimed at the visual effects and rendering market, where holding 3D models in main memory is a key aspect of workflow speed as well as full-scene production. Alongside the memory capacity difference, having double the PCIe 4.0 lanes means more access to offload hardware or additional fast storage, also important tools in the visual effects space. Threadripper Pro falls very much into the bucket of 'if you need it, this is the option to go for'."
Amúgy játékra se ideális de azért a 64 magos EPYC -et lenyomja ..
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz TESCO-Zsömle #4768 üzenetére
>> Amúgy játékra se ideális de azért a 64 magos EPYC -et lenyomja ..
> Azt elolvastad, hogy a linkelt dián CPU-n renderelik a Crysis-t?persze, és egy kicsit irónikusnak is szántam, de akkor ez nem teljesen sikerült ..
elnézést a vicceskedésért ...amúgy engem főleg ilyenek érdekelnek:
Achieving 11M IOPS & 66 GB/s IO on a Single ThreadRipper Workstation ( 2021-01-29)
https://tanelpoder.com/posts/11m-iops-with-10-ssds-on-amd-threadripper-pro-workstation/
TL;DR Modern disks are so fast that system performance bottleneck shifts to RAM access and CPU. With up to 64 cores, PCIe 4.0 and 8 memory channels, even a single-socket AMD ThreadRipper Pro workstation makes a hell of a powerful machine - if you do it right!
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz Petykemano #4750 üzenetére
> Ezek eléggé ilyen pletykaszintű, meg nem erősített infók.
> - zen4: +25% IPC, summa +40% perf all corekiraktak egy pontosítást is
"Author’s Note (2/8/21)
There was a miscommunication between myself and the editor with regards to the performance claims of Zen 4. The only thing that we are claiming with any level of certainty is the claim of a Genoa ES sample being 29% faster then Milan at the same clocks."És újrafeldolgozta a hírt a wccftech is ..
- megismételve a ChipsandCheese -es hírt .. + körítés
- meg, hogy eredetileg is az AMD a ZEN4/ZEN5 -re fókuszált az újabb nagy áttöréssel.
"We are focusing on Zen 4 and Zen 5 to be extremely competitive."Mottó: "A verseny jó!"
-
S_x96x_S
őstag
> a teljes vízhűtés, ...
> Ha ebből csak május környékén lesz kézzel fogható CPU ..pár WRX80 -as új alaplapot márciusra jelezték ..
"The ASUS Pro WS WRX80E-SAGE SE Wi-Fi is expected to be available in North America from March."
ami kvázi egy 8 csatornás workstation-os epyc .. magasabb frekvenciával..ha nincs még alaplapod, akkor érdemes lehet ezt is mérlegelni.
amúgy szerintem kicsi az esély - hogy márciusban hozzá lehessen jutni valami 8 csatornás ZEN3 -as cucchoz - amíg a " ABF substrate shortages" probléma meg nem oldódik ..
utána is kell 1-2 hónap míg az elmaradt megrendeléseket ki nem egyenlítik.de azért reménykedek
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
XT verzió ?
AMD Allegedly Launching Ryzen 5000 Zen 3 Desktop CPU With 5 GHz+ Clocks, Announcement Planned For 15th February
https://wccftech.com/amd-ryzen-5000-zen-3-desktop-cpu-with-5-ghz-clocks-announcement-next-week-rumor/
https://twitter.com/1usmus/status/1360168025794691073Mottó: "A verseny jó!"
-
S_x96x_S
őstag
> Az a helyzet, hogy a Threadripper nálam elásta magát
az első benyomás fontos ..
de tekintheted úgy is, hogy ez már egy új socket.
vagyis várd meg a teszteket .. aztán döntesz ...amúgy én most ilyet hajtok .. ( persze csak bérlés, Google Cloud )
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
Address sizes: 48 bits physical, 48 bits virtual
CPU(s): 8
On-line CPU(s) list: 0-7
Thread(s) per core: 2
Core(s) per socket: 4
Socket(s): 1
NUMA node(s): 1
Vendor ID: AuthenticAMD
CPU family: 23
Model: 49
Model name: AMD EPYC 7B12
Stepping: 0
CPU MHz: 2250.000
BogoMIPS: 4500.00
Hypervisor vendor: KVM
Virtualization type: full
L1d cache: 128 KiB
L1i cache: 128 KiB
L2 cache: 2 MiB
L3 cache: 16 MiB
NUMA node0 CPU(s): 0-7
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; Full AMD retpoline, IBPB conditional, IBRS_FW, STIBP conditional, RSB filling
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Not affected
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good no
pl nonstop_tsc cpuid extd_apicid tsc_known_freq pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy cr8_legacy
abm sse4a misalignsse 3dnowprefetch osvw topoext ssbd ibrs ibpb stibp vmmcall fsgsbase tsc_adjust bmi1 avx2 smep bmi2 rdseed adx smap clflushopt clwb sha_ni xsaveopt xsa
vec xgetbv1 clzero xsaveerptr arat npt nrip_save umip rdpid
> Amúgy nézegettem én is ezeket a WRX80 WEPYC alaplapokat,
meg kell várni a ZEN-4-et .. hátha akkora kiforrja magát ..Nekem a legnagyobb bajom, hogy mindjárt itt a PCIe5.0 , ami az alacsony latency-jával új forradalmat indít el .. és a PCIe4.0 már lassan elavult lesz.
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz Petykemano #4784 üzenetére
> De sajnos a "bejelentés" nem az AMD-től van, hanem a Ryzen OC/UV tool készítőjétől.
hát ezt benéztem ...
elnézést ..
( A fránya wccftech megvezetett - persze ez inkább tanulság és nem mentség .. )
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
> CPU MHz: 3802.226
> BogoMIPS: 4591.33és az én ( bérelt ) ZEN2 -es BogoMIPS-em kevesebb ... ( csak 4500.00 )
amúgy rákerestem , és jó tudni, hogy létezik barkács EPYC overlock ;
ma is tanultam valamit!Mottó: "A verseny jó!"
-
S_x96x_S
őstag
A legfrissebb ( ~ 1 hetes : Tue 09 Feb 2021 02:03:38 AM CET )
corporate AMD pdf szerint
- A ZEN4 - 5nm -en várható. és mintha 2022 előtt ?? lenne várható .. ( slide 28 )
- 3rd GEN ZEN3 AMD EPYC - Launch Q1 2021 ( március ? slide 25 )Mottó: "A verseny jó!"
-
S_x96x_S
őstag
February 16,
"Lenovo ThinkStation P620 Review: A Vehicle for Threadripper Pro"
https://www.anandtech.com/print/16482/lenovo-thinkstation-p620-review-a-vehicle-for-threadripper-provan benne egy érdekes benchmark : "8-Channel Benefits over 2-Channel"
Érdekesek az upgrade árak - szokni kell őket.
- 64-core TR Pro 3995WX for +$12200,
- 8x64 GB of DDR4-3200 RDIMM ECC (512 GB total) for +$12000,
- dual NVIDIA Quadro GV100 32 GB graphics cards for +$39650,persze ez lehet, hogy valami sales trükk .. mert sok mindenre 42% diszkont van.
A full config "$60634" helyett akciósan csak "($35168)"
minimum config "$3619" helyett "($2099)"
Persze márciusban talán ZEN3-as TR upgrade várható. szóval hamar elavul ...Mottó: "A verseny jó!"
-
S_x96x_S
őstag
új memória típus: HBM-PIM ;
"The new processing-in-memory (PIM) architecture brings powerful AI computing capabilities inside high-performance memory, to accelerate large-scale processing in data centers, high performance computing (HPC) systems and AI-enabled mobile applications.
..
Instead, the HBM-PIM brings processing power directly to where the data is stored by placing a DRAM-optimized AI engine inside each memory bank — a storage sub-unit — enabling parallel processing and minimizing data movement. When applied to Samsung’s existing HBM2 Aquabolt solution, the new architecture is able to deliver over twice the system performance while reducing energy consumption by more than 70%. The HBM-PIM also does not require any hardware or software changes, allowing faster integration into existing systems."
Samsung Develops Industry’s First High Bandwidth Memory with AI Processing PowerMottó: "A verseny jó!"
-
S_x96x_S
őstag
"The 48-core AMD EPYC 7643 Zen 3 server processor has made another appearance on Geekbench, turning up on version 5 of the synthetic benchmark. The Milan chip pulled out a gigantic multi-core score of 33,348 points that puts it +33% ahead of the average result managed by the 64-core AMD Ryzen Threadripper 3990X."
( via notebookcheck-feb-17)Azért az új 28 és 56 magos ZEN3-as EPYC-et szoknom kell még
... lehet, hogy a marketing kérte?
( hogy tudják benchmarkolni az Inteles 28C -al ? azonos magszám ? )
- EPYC 7453 - 28C 2.40GHz 180W vs. XEON GOLD 6348 28C 2.80GHz 235W
- EPYC 7663 - 56C 2.10GHz 225W
https://twitter.com/momomo_us/status/1360638364823801857vagyis most az új EPYC magszám variációk - a momomo_us lista alapján:
- 8, 16, 24, 28, 32, 48, 56, 64
Az Intel "core" Skála
- 8, 12, 16, 18, 20, 24, 26, 28, 32, 36, 38, 40Mottó: "A verseny jó!"
-
S_x96x_S
őstag
INTEL Alder Lake-S: December 2021 ( Gen5, DDR5, ... )
AMD AM5 : csak jövőre lesz DDR5 ...
"Intel Alder Lake-S will require new motherboards with an LGA1700 socket. This is necessary as the platform shifts to DDR5 and PCI Gen5 interfaces. The competition from AMD may be slightly delayed till 2022, the site further claims. The AM5 based motherboards would too support the DDR5 memory."
( via videocardz )Mottó: "A verseny jó!"
-
S_x96x_S
őstag
érdekesség ..
"Jim Keller: The Future of Computing, AI, Life, and Consciousness | Lex Fridman Podcast"
https://www.youtube.com/watch?v=G4hL5Om4IJ4&list=WL&index=1 ( ARM vs. INTEL )Mottó: "A verseny jó!"
-
S_x96x_S
őstag
TSMC .. ( gyártás - digitimes hirek )
- A TSMC mást prioritizál --> még pár hónapig ellátási probléma lesz az AMD-s chipekből
- Lehet, hogy lesz az EU-ban is TSMC gyár
- Politikai nyomás a TSMC-re, hogy az USA -ban is csináljon gyárat ..Friday 19 February 2021
"TSMC has agreed to give top priority to orders placed by its automotive chip clients, giving them SHR (super hot run) status, according to sources at Taiwan-based analog IC firms."Thursday 18 February 2021
"TSMC could set up an advanced wafer fab in Europe, where demand for leading-edge chip fabrication is promising, according to industry observers."Monday 8 February 2021
"TSMC is believed to have been under political pressure to set up an advanced wafer fab locally in the US in the wake of the US-China trade war. The pure-play foundry has also come under the spotlight amid a global shortage of automotive chips, according..."Mottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz Petykemano #4814 üzenetére
> Ez vajon miért nem Arm alapon készült?
piaci megfontolások?
- konkrétan a legtöbb játék még X86-os. ( Steam ...)
- az Apple meg nem adja ki másoknak az M1-et.
- az ARM-es Androidos játékokat meg a saját telefonon is lehet futtatni.De a kérdés jogos, ha lesz tömegesen elérhető M1 kategóriájú ARM-es proci,
akkor biztosan próbálkoznak.Mottó: "A verseny jó!"
-
S_x96x_S
őstag
A Xilinx ( az AMD leendő része ) - egy 16 magos ARM-et kombinált FPGA-val.
szerintem csak nem adják fel az ARM-es vonalat - a teljes egyesülés után se ...
Xilinx pops a 16-core 64-bit Arm system-on-chip from NXP into its latest FPGA-based 100Gbps smart NIC https://www.theregister.com/2021/02/23/xilinx_sn1000_launch/"Speaking of hardware, what's in the Alveo SN1000? Specifically, Xilinx is kicking off with the SN1022, a full-height, half-length PCIe 4 card due to go on general sale in March. This includes a TSMC-fabbed 16nm UltraScale+ XCU26 FPGA, an NXP processor with 16 64-bit Cortex-A72 cores running at 2GHz with 8MB of cache – nothing to sniff at but certainly not the top of end of Arm's range – and 4GB of DDR4-2400 RAM for the 16nm NXP LX2162A system-on-chip and 8GB of the same for the FPGA. The XCU26 is said to have a million LUTs."
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
> Ez mennyire jó/hasznos termék ?
spéci ..
Cloud / Enterprise termék ..
programozható - rekonfigurálható SmartNIC ..A Xilinx nem régen lépett be erre a piacra ..
2020 Március "Xilinx Enters The SmartNIC Market"
Ahol hálózaton keresztül nagytömegű adat jön be és fontos a kicsi késleltetés és a gyors adatfeldolgozás .. előszűrés ..
.. ott életmentő ...Lényegében a CPU által elvégzett feladatokat tolja át egy spéci kártyára.
- egyedi videó kódolás - dekódolás ..
- elektronikus kereskedelem ( gyors válaszidő .. )
- bejövő hálózati forgalom biztonsági elemzése .. ( DDOS attack kivédése )
- spéci titkosítás ..
- vagy gondolj az 5G -re .. rengeteg adat ..
- ...
de lényegében ABU is hasonlót irt ..
https://prohardver.hu/hir/velemeny_nem_fpga_lenyeg_amd_xilinx.htmlez hasonló mint a PS5 -nek az ssd-be épített célprocija .. csak hálózatra és újraprogramozható.
A PS5-nél A "Kraken (and ZLib) " tömörítést végző co-proci
9 ZEN2-es CPU mag munkáját váltja ki !És egy szerveren ..
már egy-két 100Gpbs -es folyamatos adatkapcsolat is már akkora adatmennyiséget jelenthet
.. hogy egy normál 64 magos szerveren is érezhető ...
itt ( ~ tippelek ) akár 8 - 16 magot is kiválthat - tehermentesíthet .. egy SmartNICde ez csak az én ködösen naiv véleményem .. nem vagyok annyira járatos benne.
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz Devid_81 #4823 üzenetére
akkor a tweetből - összefoglalva:
AMD Genoa Epyc - SP5 / ZEN4 :
- 96-cores (192 threads)
- 12-channel DDR5-5200
- 128 PCIe 5.0 lanes (160 for 2P)
- 320W TDP (cTDP 400W)
- SP5 (LGA-6096) socketDe akkor mi várható az AM5 -re?
> Azt hittem 12 magos lesz egy CCD, de allitolag csak tobb CCD lesz 12x8
valószínűleg az üzleti racionalitás miatt maradt 8;
- mert desktop fronton még mindig ennek van értelme
és így lehet több helyen is felhasználni a chipletet ...- meg hátha jön az AVX-512 / DL ... és emiatt több tranzisztor kell egy magra.
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
> Elég sokan vesznek 12 magosat is,
szerintem
- nincs erőforrás 2 különböző chiplet gyártására
( 8 és 12 magos )
- egyfajta chiplettel alacsonyabban lehet tartani a költségeket
( gyártási / tervezési / tesztelési)
- és a keresleti piac nagy része is R3 / R5 lesz ;
( vagyis 4, 6 , 8 -mag)> Ez nekem inkább úgy hangzik, hogy minek többet adni, mivel így sincs konkurencia.
gazdaságosan is lehet többet adni , ahogy eddig is történt ( 12 és 16 mag )
vagyis a költségek alacsonyan tartásával.Mottó: "A verseny jó!"
-
S_x96x_S
őstag
> Elég pici lesz 5 nm-en a 8 magos CCD, ...
> Kivéve, ha megint duplázzák a cache-t. ...kérdés, hogy az AVX-512 -vel mi lesz.
Az Intel most gőzerővel beindult .. és mindenhova, még az i3-as procikba is berakja ..
És ha akar AVX-512 -öt az AMD,
akkor lehet, hogy duplázni kell megint a lebegőpontos dolgokat, és akkor már 12 mag nem könnyen fér el ..> Pedig örültem volna egy 24 magosnak, TR-t nem szeretnék megint, túl batár.
AM5-ben?
ha marad a 2 csatornás memória ( mégha DDR5-is ), akkor az lesz a szűk keresztmetszet.
( habár CXL 2.0-ban már mindenféle memória-kütyüket is rá lehet kötni majd )Ha igaz, ami ABU -írt .. és lesz kicsi epyc .. akkor lehet, hogy lesz kics TR is?
és akkor a te problémád is megoldva.
Vagyis nem az AM5 lenne kibővítve 24 mag-ig , hanem az epyc(TR?) lenne lehozva a 16c-24c -es területre.Abu85(2020-08-27): " ... Az egész szerverpiacot figyelembe véve a tipikusan keresett teljesítményszint 12-24 mag körüli.
...
Az ötlet szerint az EPYC mellé jönnének a kicsi EPYC-ek. És ezen konkrétan azt kell érteni, hogy kisebb tokozást kapnának, nyilván a csúcsmodellnek is visszafogott teljesítménye lenne, de eközben megmaradna minden olyan szerverfunkció, ami a nagyobb testvérek része, tehát igazi szervereket lehetne építeni rájuk.
...
Az útiterv szerint ennek két lényeges haszna van. Egyrészt a nagy EPYC-ek árszintet válthatnak. Ha valakinek kell a legjobb teljesítmény, illetve a legjobb hatékonyság, akkor fizesse meg. Viszont az biztos, hogy a tömegpiacnak nem fog kelleni, na ez a réteg ne fizessen sokat a szerverért, egyszerűen vegye meg a kívánt teljesítményt jelentősen a megszokott szerverpiaci árak alatt.
Az AMD valószínűleg abból indul ki, hogy ha a tipikus igények jövőre csak 32 magig nőnek meg, akkor is ki tudják azokat szolgálni egy asztalinál alig komplexebb rendszerrel, amelyet pusztán ki kellene egészíteni a szerverszintű funkciókkal. Erre a Zen 4 chiplet ideális lehet, és egy ilyen csomag ára egészen alacsony lehetne, miközben tudásban semmivel sem kevesebb a nagyobbik dizájnnál. A combosabb EPYC nyugodtan felveheti a fonalat például 32-magos szinten, és lerakhatja akár 100 mag felett, ez részletkérdés a tömegpiac megnyerése szempontjából. Így maga a szerverek szegmense szegmentálódna az AMD konstrukciójában, és a nagy dizájnoknak valós előnye lenne az árukhoz mérten, hiszen ha kiderül, hogy mégse elég mondjuk 32 mag, akkor processzorcserével lehet feljebb menni. Ez akár tervezhető is a vásárlásnál, szemben a kisebb dizájnokkal, ahol alapvetően 32 mag lehetne a felső határ, cserébe nagyságrendekkel kevesebb pénzt kellene kiadni értük."
https://prohardver.hu/hir/amd_kicsi_epyc_tamadna_be_szerverpiac.htmlMottó: "A verseny jó!"
-
S_x96x_S
őstag
>Azért az a DDR5 közel dupla sávszélességet jelent,
igen - "elméletben" ( az elméleti maximimum 2x)
DDR5-3200 ~= DDR4-3200 ( Peak BW GB/s = 25 )
most vagyunk a DDR4 csúcsán - a végén .. és a DDR5 technológia elején.ha a ddr5 életciklus elején ~1,4x növekedéssel számolunk memória sávszélességnél
( az első években .. az életciklus végén majd meglesz a 2x )
akkor 1.4x proci teljesítmény növekedésnél még mindig ugyanott tartunk.
( nagyjából és erős egyszerűsítéseket alkalmazva )
De a te javaslatodnál:
- növekszik a magszám ( 1,5x )
- növekszik a magonkénti teljesítmény ; zen3- > zen4 ( ~ 1,15x )
vagyis 1,5x * 1,15x = 1,72 !Az 1,15x ( 15% ) -ot elég konzervatívra vettem .. írhattam volna akár 1.4x ( +40%) -et is
2021-02-09: "Incredible rumors concerning AMD’s Zen 4 and Zen 5 microarchitectures have been shared that suggest impressive generational gains. Zen 4 will apparently offer +25% IPC gains, +40% overall performance, and it may also bring all-core processing at 5 GHz. " [link]Mottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz Petykemano #4834 üzenetére
> zen4 + AVX3-512 + bfloat16
ez nem rossz .. kell a HPC-hez is és az Intel marketinges érvek egyrészét is hárítja.
akkor már csak az új AMX ( Advanced Matrix Extensions ) van hátra.> különös ez a sok új info a zen4-ről és a Genoa-ról. Még a Milan se rajtolt el.
sok oka lehet ..
de az én kedvenc magyarázatom:
- A háttérben az AMD is gőzerővel fejleszt ( kevésbé marketingelve mint az Intel )
megint sok mindent be szeretnének tolni .. de kockázatos a roadmap.
vagyis csak a legvégén derül ki, hogy mi kerül a vásárlók asztalára.
( és most ért ki a tesztelőkhöz a teszt példány .. és mindenki szivárogtat egy picit )
és az Intel mostani információ dömpingjét próbálják ellensúlyozni- Ráadásul a TSMC is növeli a kapacitását 2021H2 -re.
"TSMC is poised to scale up its 5nm process manufacturing capacity to 105,000 wafers monthly in the first half of 2021, up from 90,000 units in fourth-quarter 2020, with plans to further expand the process capacity to 120,000 units in the second half..."
https://www.digitimes.com/news/a20210226PD207.htmlNem kérdés, hogy az AMD a második lesz a DDR5/PCIe5.0 bevezetésében ..
( legalábbis marketing és ígéretek terén )
de az már fontos, hogy mennyivel lemaradva!A 2022-es CES nagyrészt erről fog szólni ..
rengeteg ( DDR4/Gen5 ) termékeket fognak bemutatni
és az AMD-nek is ott kell lennie - nyilvános roadmappel és várható teljesítmény növekedésselpersze 1 év múlva ( 2022március) már látjuk,
hogy az AMD vagy az Intel ( Gen4/ddr5 ) bevezetése lett jobb ..
mit mutatnak a tesztek, stb.
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
AMD Ryzen Threadripper Pro Hits Retail Availability and an ASUS Update
https://www.servethehome.com/amd-threadripper-pro-hits-retail-asus-update/
( WEPYC )Velocity Micro ProMagix HD150 WRX80 TR Pro System
" One of the other nice features is that the ProMagix HD150 starts at under $4,500 which is a lot lower than Lenovo has been selling the P620 for while it had little to no competition."
lehet követni a WRX80 elérhetőségét az EU-ban
https://geizhals.de/?fs=WRX80&hloc=at&hloc=de&hloc=eu&hloc=pl&hloc=uk&in=
pl.
AMD Ryzen Threadripper PRO 3955WX, 16C/32T, 3.90-4.30GHz, boxed ~= 1100 EUR
AMD Ryzen Threadripper PRO 3975WX, 32C/64T, 3.50-4.20GHz, boxed ~= 2700 EUR
AMD Ryzen Threadripper PRO 3995WX, 64C/128T, 2.70-4.20GHz, boxed ~= 5500 EUR
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
Konkurencia vs. AMD
"Intel Core i7-11700K Review: Blasting Off with Rocket Lake"
https://www.anandtech.com/print/16535/intel-core-i7-11700k-review-blasting-off-with-rocket-lake
Összefoglaló:
"Our results clearly show that Intel’s performance, while substantial, still trails its main competitor, AMD. In a core-for-core comparison, Intel is slightly slower and a lot more inefficient. The smart money would be to get the AMD processor. However, due to high demand and prioritizing commercial and enterprise contracts, the only parts readily available on retail shelves right now are from Intel. Any user looking to buy or build a PC today has to dodge, duck, dip, dive and dodge their way to find one for sale, and also hope that it is not at a vastly inflated price. The less stressful solution would be to buy Intel, and use Intel’s latest platform in Rocket Lake."
INTEL - Játékok:
"Users looking at our gaming results will undoubtedly be disappointed. The improvements Intel has made to its processor seem to do very little in our gaming tests, and in a lot of cases, we see performance regressions rather than improvements. If Intel is promoting +19% IPC, then why is gaming so adversely affected?"Mottó: "A verseny jó!"
-
S_x96x_S
őstag
válasz solfilo #4839 üzenetére
> Eléggé meglepett, főleg a játékos rész.
még változhat, és egy nagy bios/mikrokód frissítés várható
és a pletykák szerint ezért tették későbbre a hivatalos megjelenést és a "review"-ket
( ---> március 30-ra )"Earlier rumors have suggested that Intel is still in a process of polishing the microcode and that’s the reason why launch reviews are also planned for March 30th. This BIOS update may have an impact on performance figures scrupulously collected by AnandTech. Whether that’s the case, we will see in a month’s time when their launch review goes live."
https://videocardz.com/newz/intel-core-i7-11700k-retail-cpu-tested-in-the-first-comprehensive-reviewMottó: "A verseny jó!"
-
S_x96x_S
őstag
Március 15 - EPYC 3rd gen launch --> https://www.amd.com/en/events/epyc
persze ezt logikus ..
és ha az olasz zászlót 90fokkal elfordítjuk -> rögtön a magyar zászlót kapjuk ..
vagyis emiatt lett Március15 !de máshol is helyeztek el üzeneteket, pl. egy tejesdobozon.
( via STH )
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
új támadási mód .. persze elsőként az Inteles procikon tesztelik,
Lord of the Ring(s): Side Channel Attacks on the
CPU On-Chip Ring Interconnect Are Practical
https://arxiv.org/abs/2103.03443
"To address the first challenge, we perform a thorough reverse engineering of Intel’s “sophisticated ring protocol” [60,90] that handles communication on the ring intercon-
nect."
....
"Finally, AMD CPUs utilize other proprietary technologies known as Infinity Fabric/Architecture for their on-chip interconnect [20, 96]. Investigating the feasibility of our attack on these platforms requires future work. However, the techniques we use to build
our contention model can be applied on these platforms too."Mottó: "A verseny jó!"
-
S_x96x_S
őstag
piac/konkurencia: RIP MIPS
A MIPS bedobta a törölközőt és dobta a saját CPU architektúráját.
EZentúl RISC-V -t tervez.
Egyperces csenddel emlékezzünk meg erről a régi nagy múltú CPU architektúráról ..vége:
- https://en.wikipedia.org/wiki/MIPS_architecture
- https://hu.wikipedia.org/wiki/MIPS_architekt%C3%BAra"Wait, What? MIPS Becomes RISC-V
Classic CPU Company Exits Bankruptcy, Throws in the Towel"
https://www.eejournal.com/article/wait-what-mips-becomes-risc-v/
"What a long, strange trip it’s been. MIPS Technologies no longer designs MIPS processors. Instead, it’s joined the RISC-V camp, abandoning its eponymous architecture for one that has strong historical and technical ties. The move apparently heralds the end of the road for MIPS as a CPU family, and a further (slight) diminution in the variety of processors available. It’s the final arc of an architecture. "
persze 3 éve még :
2018: ABU: MIPS magokat használó rendszerchippel készül az Intel
https://prohardver.hu/hir/intel_mips_magokat_hasznalo_rendszerchip_keszul.html
2018-dec: "Nyílt forráskódú lesz a MIPS"
https://prohardver.hu/hir/nyilt_forraskodu_lesz_mips.htmlvagyis a RISC-V architektúrára sikeresen kinyirta a MIPS-et.
... ennyi .. vége ...Mottó: "A verseny jó!"
-
S_x96x_S
őstag
mivel a Xilinx az AMD része lesz - esetleg érdekes lehet az Intel OneAPI - jára adandó válaszként értelmezni.
A jövő programozói már nem CPU -ra, hanem CPU+GPU+FPGA -ra irnak kódot.FPGA: A Xilinx megnyitotta a "Vitis HLS Tool" forráskódját ..
Xilinx Opens Up Vitis HLS Tool for FPGAs
https://www.eetimes.com/xilinx-opens-up-vitis-hls-tool-for-fpgas/
HN: https://news.ycombinator.com/item?id=26418307
https://github.com/Xilinx/HLS"The Vitis™ HLS is a high-level synthesis tool that allows C, C++, and OpenCL™ functions to become hardwired onto the device logic fabric and RAM/DSP blocks. It implements hardware kernels in the Vitis application acceleration development flow; and to use C/C++ code for developing RTL IP for FPGA designs in the Vivado® Design Suite."
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
> Am4-ről meg ráérős váltani, mert 16 magig van lehetőség a jövőben fejleszteni.
> így egy + 5-10 évig még el leszünk vele szerintembár teljesen egyetértek abban,
hogy a következő 2 évben a "gyakorlatban"
nem sok értelme lesz egy átlag felhasználónak..de azért az AM5 ( Gen5, DDR5 ) -nek is meglesz az előnye.
Igazából a CXL-es latency csökkentés lehet a "rögtön" kézelfogható előny a játékosoknak.
( feltéve ha a GPU is támogatja )itt már extrém <10ns alattiról is írnak .. habár ez már eltér a szabványtól ..
"The XpressConnect family delivers extended reach at >80% lower latency than the PCIe specification, with a pin-to-pin latency of <10 nanoseconds." [link]összehasonlításnak:
"HPC-oriented Latency Numbers Every Programmer Should Know"
https://gist.github.com/eshelman/343a1c46cb3fba142c1afdcdeec17646Amit feltételezek, hogy az Intel az új GPU .ját és az új CPU -ját ( Gen5 , CXL -el )
( ~ 1 év múlva .. jövőre )
már valami csoda új - direkt erre fejlesztett játékokkal demózza ..
mert valamivel "rögtön el kell adni az új hardvert"
és mivel lehet legjobban demózni ?
Hát játékkal ... főleg ha a saját hardveredre van optimalizálva.
( legalábbis én ha Inteles menedzser lennék .. erősen elgondolkodnék ezen )De az is igaz, hogy az új platform - újabb problémákkal jár ..
és aki az elején beleugrik .. eléggé "béta-terméket" kap ..
Viszont manapság már sokaknak ez a ( tuning, hibakeresés, optimalizáció )
több izgalmat okoz .. mint a kiszámítható játékok ..mindenki majd dönt a pénztárcájával ..
Nekem azért izgalmas ez a CXL , PCIe5 váltás .. sok új lehetőséggel.
Mottó: "A verseny jó!"
-
S_x96x_S
őstag
Nem elfelejteni ..
Ma 16:00
"Introducing 3rd Gen AMD EPYC Processors for the Modern Data Center"
https://www.youtube.com/watch?v=xtrhHH0kQI0"Join CEO Dr. Lisa Su, CTO Mark Papermaster, Senior VP and GM of Datacenter and Embedded Solutions Business Group, Forrest Norrod, Senior VP and GM of Server Business Unit, Dan McNamara, and appearances by industry-leading data center strategic partners and customers in this digital launch of the 3rd Gen AMD EPYC™ Processors."
Lesz ma (ZEN3)Threadripper bejelentése .. vagy sem ?
Remélem igen !!![ Szerkesztve ]
Mottó: "A verseny jó!"
Új hozzászólás Aktív témák
- PlayStation 4/Slim/Pro konzolok teljes karbantartása, pasztázással Thermal Grizzly 1 órán belül!!
- XBOX360/One/One S/One X konzolok teljes karbantartása, pasztázással Thermal Grizzly 1 órán belül!
- XBOX Series S / X HDMI csatlakozó csere, 1 napos határidővel, 3 hónap garanciával, üzletben!!
- XBOX ONE S/X HDMI IC csere aznapi elkészüléssel, 3 hó garanciával, üzletben!
- PlayStation 5 / PlayStation 4 Fat/Slim/Pro HDMI IC / csatlakozó szerviz, garanciával, üzletből!