- Joined
- Jan 8, 2017
- Messages
- 9,404 (3.29/day)
System Name | Good enough |
---|---|
Processor | AMD Ryzen R9 7900 - Alphacool Eisblock XPX Aurora Edge |
Motherboard | ASRock B650 Pro RS |
Cooling | 2x 360mm NexXxoS ST30 X-Flow, 1x 360mm NexXxoS ST30, 1x 240mm NexXxoS ST30 |
Memory | 32GB - FURY Beast RGB 5600 Mhz |
Video Card(s) | Sapphire RX 7900 XT - Alphacool Eisblock Aurora |
Storage | 1x Kingston KC3000 1TB 1x Kingston A2000 1TB, 1x Samsung 850 EVO 250GB , 1x Samsung 860 EVO 500GB |
Display(s) | LG UltraGear 32GN650-B + 4K Samsung TV |
Case | Phanteks NV7 |
Power Supply | GPS-750C |
GA102 was faster than Navi21 most of the time and rarely on par, which is exactly what you'd expect looking at the specifications of each chip.My expectation was a repeat of Navi 21 vs. GA102, so the very least on par with AD102 with the usual pitfalls that renders it the runner-up.
And it's the same story with Navi31 and AD102, 61 TFLOPs vs 82 TFLOPs, they're both dual-issue architectures, so roughly 35% difference in pure compute performance, TPU puts the 4090 at 22% faster in raster. So, as I said, more or less what to be expected.
As you know the reference model does not consume more power than the FE 4090. And who is to say it's artificially limited by the 2x8 pin, I see no real evidence for that. Most AIBs that ship with 3x8 pin do not overclock any better than reference models, it wouldn't make sense anyway because they all have the same +15% power limit limitation as far as I know.As it stands, the 7900 XTX consumes more power than the RTX 4090 (provided you don't buy an artificially limited, 2x 8-pin model, by the way - I can make the argument that my 3090 "only" uses 375W too, but you are leaving performance on the table), and performs substantially worse than it does in its already hilariously cut down configuration.
Last edited: