But problem is Nvidia fans want AMD launching GPU with faster RT hardware support but ultimately is want Nvidia to drop price so that they can buy cheaper Nvidia GPU while still not supporting AMD ar??Source: https://wccftech.com/nvidia-blackwe...-significant-improvement-cache-sm-throughput/
GPU NAME GB202 (Blackwell, RTX5000) AD102 (Lovelace, RTX4000) GPC 12 (Per GPU)? 12 (Per GPU) TPC 8 (Per GPC)? 6 (Per GPC) SM 2 (Per TPC)? 2 (Per TPC) Total SMs 192? (+33%) 144 Sub-Core TBD 4 (Per SM) FP32 128 (Per SM)? 128 (Per SM) FP32+INT32 TBD 192 (Per SM) CUDA Cores 24,567? (+33%) 18,432 Warps TBD 64 (Per SM) Threads TBD 2048 (Per SM)
GENERATION BLACKWELL ADA LOVELACE AMPERE TURING PASCAL Process Node TSMC 3nm?5nm+ (4NP)?TSMC 5nm (4N) Samsung 8nm TSMC 12nm TSMC 16nm Launch Year 2024 2022 2020 2018 2016 Ultra-Enthusiast SKU GB202 AD102 (RTX4090) GA102 TU102 GP102 Enthusiast SKU GB203 AD103 (RTX4080, 4070Ti Super) GA102 TU104 GP104 High-End SKU GB205 AD104 (RTX4070) GA104 TU106 GP104 Mainstream SKU GB206 AD106 (RTX4060) GA106 TU106 GP106 Entry-Level SKU GB207 AD107 (RTX4050) GA107 TU116/117 GP107
Based on the article above, here is my opinion:
Main point: still manufactured on 5nm TSMC fab process, but enhanced.
Key improvement: 30% more transistors, with my estimation +15% more power (yikes!)
Result: a "Super" upgrade, with Displayport 2.1
Example:
RTX5090 = RTX4090 Super
RTX5080 = RTX4080 Super Super,
RTX5070 = RTX4070 TI Super.
IMO, since AMD have poor raytracing performance, RTX5000 will be priced:
+30% above RTX4000 series in first year of launch because it will be selling concurrently with existing stock of RTX4000 series. Then price cut in 2nd year to match RTX4000 pricing after the old stock are sold out.
![]()
The key to lower pricing is AMD launching GPU with faster RT hardware support.
Nope. I avoided melting power connector problem... and still happily using.. fact is I hardly got chance to render any Ray Tracing games also..Regret buying AMD gpu now?![]()
May I ask which cable should I use? The 12VHPWR that comes with the PSU or the 12V-2x6 adapter that comes with the GPU?hopefully 12v-2x6 wont melt with 600w, but a lot people havent got their card yet so only time will tell
Watching all the 5090/80 beta-testers be like .......
![]()
I keep a list of games I cannot play on my current rig so I might remember to play them when I upgrade graphics card.Yeah.. just make do with whatever you have on hand... still using AMD 7900XTX.... and I dun even bother buying CyberPunk.. which I feel is basically a tech demo by Nvidia...![]()

I can play them but no motivation to play or no motivation to actually buy them as price of games keep increasing also...I keep a list of games I cannot play on my current rig so I might remember to play them when I upgrade graphics card.![]()
Of cos. Astral 2.7k. 2k cheap cheap. Some more they going kenna tariffs. No more 2k for them.5080 SGD2k still got people buy?
Hahaha maybe me? But no stock until maybe... 6-7 Feb when MSI officially do stock in worldwide (MSI officially push back sales to 6 Feb), my current supplier telling me that he's hunting one that will fit in my SFF PC, NR200P Max... have to upgrade PSU from 850W to 1000W also... after that my 4090 will go to my wife for new build. I not buying from Aftershock because I dealt with their customer service previously and I had a terrible time, despite buying at the highest end, they treat their customer like ****. So I changed vendor and quite happy with them.Still waiting for the first dude here to show off his shiny 5090 after paying over 5k for it purely for gaming purposes. Wanna congratulate him.
I dont care about Nvidia dropping price tbh - I just want to see local prices make senseBut problem is Nvidia fans want AMD launching GPU with faster RT hardware support but ultimately is want Nvidia to drop price so that they can buy cheaper Nvidia GPU while still not supporting AMD ar??

Bruh. Your use case is so unique and niche. I enjoy seeing folks like you putting the AI Prowess to good use.Hahaha maybe me? But no stock until maybe... 6-7 Feb when MSI officially do stock in worldwide (MSI officially push back sales to 6 Feb), my current supplier telling me that he's hunting one that will fit in my SFF PC, NR200P Max... have to upgrade PSU from 850W to 1000W also... after that my 4090 will go to my wife for new build. I not buying from Aftershock because I dealt with their customer service previously and I had a terrible time, despite buying at the highest end, they treat their customer like ****. So I changed vendor and quite happy with them.
To be 100% fair, my current SFF build costs about 5.5k because of MSI 4090. 9800X3D (Upgraded from 7800X3D), Strix X870-I (X670E-I previously, got no clear CMOS button... ***** stupid design), 64GB 6000MHz, 2TB 980 Pro, 2TB SN850X and NR200P Max. The 5090 on its own look to cost the entirety of the best of the best build previously.
Still I'd probably upgrade since I play around with local LLMs and can't wait to play around with the various LLMs e.g. the latest DeepSeek without Quantization, Quantization drops performance by too much to make any sense on a 4090. E.g. 8Q drops token rate to 1/8, 4Q drops it to 1/4 and 2Q drops it by 1/2. If can fit into VRAM without Quantization, performance doubles or quadruples depending on how many Q you are forced to run the model previously.
While I play around with AI mostly, I also game about 5-10 hours a week. It's nice to have a GPU that can do both.Bruh. Your use case is so unique and niche. I enjoy seeing folks like you putting the AI Prowess to good use.
Only Zotac 5080s are SGD2k. Every other AIB the prices are like 2.2k up5080 SGD2k still got people buy?