- Joined
- Sep 3, 2019
- Messages
- 3,266 (1.79/day)
- Location
- Thessaloniki, Greece
System Name | PC on since Aug 2019, 1st CPU R5 3600 + ASUS ROG RX580 8GB >> MSI Gaming X RX5700XT (Jan 2020) |
---|---|
Processor | Ryzen 9 5900X (July 2022), 163W PPT limit, 80C temp limit, CO -8~12 |
Motherboard | Gigabyte X570 Aorus Pro (Rev1.0), BIOS F37h, AGESA V2 1.2.0.B |
Cooling | Arctic Liquid Freezer II 420mm Rev7 (Jan 2024) with off center mount for Ryzen, TIM: Kryonaut |
Memory | 2x16GB G.Skill Trident Z Neo GTZN (July 2022) 3600MT/s 1.38V CL16-16-16-16-32-48 1T, tRFC:280, B-die |
Video Card(s) | Sapphire Nitro+ RX 7900XTX (Dec 2023) 314~466W (366W current) PowerLimit, 1060mV, Adrenalin v24.7.1 |
Storage | Samsung NVMe: 980Pro 1TB(OS 2022), 970Pro 512GB(2019) / SATA-III: 850Pro 1TB(2015) 860Evo 1TB(2020) |
Display(s) | Dell Alienware AW3423DW 34" QD-OLED curved (1800R), 3440x1440 144Hz (max 175Hz) HDR400/1000, VRR on |
Case | None... naked on desk |
Audio Device(s) | Astro A50 headset |
Power Supply | Corsair HX750i, 80+ Platinum, 93% (250~700W), modular, single/dual rail (switch) |
Mouse | Logitech MX Master (Gen1) |
Keyboard | Logitech G15 (Gen2) w/ LCDSirReal applet |
Software | Windows 11 Home 64bit (v23H2, OSBuild 22631.4037), upgraded from Win10 to Win11 on Feb 2024 |
Not every cuda core is created equal. Its like trying to compare different cores from different CPUs.Exactly. The thing is that the raw performance uplift comes with a TDP uplift of ~equal measure, so I can't really see any perf/W increase at all.
When they announced the prices and everybody started panic selling their 2080 Ti's, I had a feeling that there was a reason behind it. Nvidia has never been famous for bringing good value propositions in the higher market segments - don't get me wrong, people, they're good cards, just not as good as Jensen claimed in the announcement. It also puzzles me how they managed to achieve a 30% performance uplift with ~double the cuda cores.
-------------------------------------------------
Its easy if you know the specifics. According to Jensen a process unit to be called cuda core must be able to execute FP32 instructions.
Pre-Turing architectures every cuda core was capable of 1 Int or 1 FP instruction, and cannot execute both.
This has changed in Turing. In Turing every cuda core was performing exclusively FP instructions and the number was 4352. For Int instructions Turing had another 4352 units that was not counted for cuda cores (see Jensen)
Ampere 3080 now has (4352x2) 8704 cuda cores and the 4352 of them are executing exclusively FP instruction and the other 4352 are executing Int or FP instructions, but not both.
See?
Last edited: