The stuff Hacker News actually recommends.

GPUs and CPUs

AMD Ryzen 9 vs NVIDIA GeForce RTX

No clear winner: owners rate them about the same.

AMD Ryzen 9NVIDIA GeForce RTX
Owners with an opinion105617
Positive verdicts83% of 7576% of 416
Positive/negative since 202440/11214/71
Happy after 3+ years110
Longest happy owner4 yrs6 yrs
Praised forperformance and reliabilityperformance and value
Criticized fornoisenoise
Newest comment

Updated

What owners wrote about both

AMD Ryzen 9: NeutralNVIDIA GeForce RTX: Neutral
> The number of people with hardware to run a model like that is very small. Not completely obvious. I have a Ryzen 7900 with 192GB of DDR5 and a 4090. Pedestrian stuff 12 months ago (around 3k€ all-in). It runs 3.8 Flash-Next at 120tok/s. https://github.com/AnnoyingTechnology/nvidia-4090-flash-next...
AMD Ryzen 9: NeutralNVIDIA GeForce RTX: Neutral
hats off -- fantastic work -- worked out the box Operating System: TUXEDO OS KDE Plasma Version: 6.6.5 KDE Frameworks Version: 6.24.0 Qt Version: 6.10.2 Kernel Version: 6.17.0-124040-tuxedo (64-bit) Graphics Platform: Wayland Processors: 24 × AMD Ryzen 9 3900 12-Core Processor Memory: 32 GiB of RAM (31.2 GiB usable) Graphics Processor: NVIDIA GeForce RTX 2070/PCIe/SSE2 (8 GiB) Manufacturer: PC Specialist LTD Product Name: NH5xAx My DeepSeek V4 can see!
AMD Ryzen 9: Likes itNVIDIA GeForce RTX: Likes it
I'm using qwen 3.6 35B unsloth 4 bit with my 5950x (128 gb memory) and a 3060 12 gb gpu with a self made harness. At 10k context I get about 40 tps generation and 500 tps prefill. At 100k context I get about 25 tps generation and 400 tps prefill. It works, but I often use gpt or claude to make a detailed enumerated plan of what I want to do first, then have qwen follow it. I'm not sure if it is economical or not, but I have solar on the roof so the power use is not really an issue and I already have the hardware. The biggest benefit for me is that it's all done locally, and I know the harness is not uploading anything or sending telemetry to someone else.
AMD Ryzen 9: Likes itNVIDIA GeForce RTX: Likes it
I got an entire Prebuilt PC when POW ended from a miner. AMD 5950x, 64gb ram, 3090, 2tb SSD, all for $1700. This was 4 years ago when the 3090 was $1500 by itself, and with minor upgrades it's still going strong today. It's wild to think that the system now is worth at least as much as I paid for it then if not much more than that. I saw a similar one going for $2500.
AMD Ryzen 9: NeutralNVIDIA GeForce RTX: Neutral
So I randomly ended up buying a lot of computer gear last year because I killed a perfectly fine 4 year old PC and couldn't decide what I wanted to replace it. I thought at the time "this is an expensive mistake" but you look at the prices I paid for parts a year ago and it's mind-blowing eg: - 4TB Samsung Pro 990 SSD for $150 (now $940) - 64GB DDR kit for a laptop $180 (now $700) - 64GB DDR5 CL30 kit for a desktop $200 (now $950) - 9800X3D/5070Ti PC $1800 - 2TB Samsung Pro 990 $95 I think? I honestly don't even remember why I bought this It's really depressing now. Normally at this point in the NViida product cycle we'd be expected a 50x0 Super series. I think it's all but confirmed we won't see those until next year. I think the 50x0 series will last a lot longer than the 40x0 series. So it's going to be interesting to see what happens when this hits phone makers who also need RAM. There certainly won't be a RAM increase this year and there'll likely be a price bump. Apple may be able to absorb this to some degree because of anyone I expect them to have long term contracts. Still, Apple has temporarily delisted the base 16GB Mac Mini and removed the 512GB Mac Studio so they aren't unaffected. But I think this SSD/RAM price hike has basically killed the Steam Machine, which is sad. Valve obviously didn't lock in long-term contracts before announcing it. Woops. The Steam Deck is also a hard find as a result. We've seen an almost unprecedented price hike on the PS5, which is a…
AMD Ryzen 9: Likes itNVIDIA GeForce RTX: Likes it
I didn't really understand the performance table until I saw the top ones were 8B models. But 5 seconds / token is quite slow yeah. I guess this is for low ram machines? I'm pretty sure my 5950x with 128 gb ram can run this faster on the CPU with some layers / prefill on the 3060 gpu I have. I also see that they claim the process is compute bound at 2 seconds/token, but that doesn't seem correct with a 3090?
AMD Ryzen 9: NeutralNVIDIA GeForce RTX: Neutral
I always find some things that doesn't work with my PC on windows 11. Sometimes things as simple as moving files in explorer makes it hangs where I had to restart explorer.exe. This is embarrassing really that windows can't get this right. There are so many times where I was frustrated and wished that I can just use my macbook pro as my only workstation. I just wish that steam on linux has full support for most games that are it supports in windows then i'll make the switch my pc is not even that old, its ryzen 9 5900x with rtx 3080 and 32gb ram. however it is sluggish compared to my m1 pro macbook pro
AMD Ryzen 9: Likes itNVIDIA GeForce RTX: Likes it
I don't, I'm a REALLY light user. smaller LLMs work pretty well. I used a 40gb LLM and it was _pokey_, but it worked, and switching them is pretty easy. This is a 12 core Xeon with 64Gb RAM...my M4 mini is....okay with smaller LLMs, I have a Ryzen 9 with a RTX3070ti that's the best of the bunch, but none of this holds a candle to people that spend real money to experiment in this field.
AMD Ryzen 9: Likes itNVIDIA GeForce RTX: Likes it
I'm having a different experience. Aquarium: 60fps until 20k fish, where I hit 50fps. 30k at 34fps Blobs: maxed out resolution and number of blobs, still 60fps Field: 60fps at "lots" Fishtank: 60fps with 1k fish and sharks Spacerocks: 60fps on lots Sprites: 60fps on 10k System: - FF 140.0.4 - Kernel: 6.12.37 - CPU: AMD Ryzen 9 5900X - NVIDIA 4080S (575.64) - 186 tabs open (mostly YouTube. >20 active) I get a bit worse on my M2 Macbook Air (128 tabs), but pretty close results. Maybe you need to open more tabs?
AMD Ryzen 9: NeutralNVIDIA GeForce RTX: Neutral
> Do you remember what Windows and Linux and MacOS were like in 90s? What exactly is the software we are comparing? Yes, yes I do. Since then the computer have become several orders of magnitude more powerful. You cannot even begin to imagine how fast and powerful our machines are. And yet nearly everything is barely capable of minimally functioning. Everything is riddled with loading screens, lost inputs, freeze frames and janky scrolling etc. etc. Even OS-level and professional software. I now have a AMD Ryzen 9 9950X3D CPU, GeForce RTX 5090 GPU, DDR5 6000MHz RAM and M.2 NVME disks. I should not even see any loading screen, or any operation taking longer than a second. And yet even Explorer manages to spend seconds before showing contents of some directories.
AMD Ryzen 9: Likes itNVIDIA GeForce RTX: Likes it
I really have no choice but to install Linux now. The nag screens in Windows 10 are already getting unbearable. The over 5 year old system is still running games at high detail just fine (3950x CPU, 2080 GPU, 32Gb ram running at 2k resolution). TPM 2.0 was finalized in Nov 2019 after the above system was built. Hardware isn't progressing that quickly anymore which makes a hardware requirement like TPM 2.0 very arbitrary honestly. Microsoft could have easily made Windows 11 work without TPM 2.0 as the workarounds attest to. It could have been "if you don't have this, you don't get some security features but otherwise it's fine". My conspiracy sense is tingling at the hard requirement for TPM 2.0 that really doesn't need to be there. A bit like when MS bought Skype and moved it from peer to peer voice to a central server based system with no monetary gain from doing that, someone obviously decided that needed to be done. We don't need to be asking people to throw away perfectly performant hardware to upgrade. There's actually nothing in Windows 11 enabled by the requirement since Windows 10 also supported TPM 2.0 already, it just didn't require it. I've also never been subject to the attacks TPM 2.0 protects against. There may be cases where TPM 2.0 is wanted, i can imagine it's useful for servers, but i really don't need it.
AMD Ryzen 9: NeutralNVIDIA GeForce RTX: Neutral
Using TheBloke nous-hermes-2-mixtral-8x7b-dpo.Q5_K_M.gguf, llam.cpp with oobabooga, on Ryzen 9 7950x and RTX 2070 super, offloading 4 layers to the GPU. 27,631MiB system memory used. 7358/8192MiB usage on the GPU. I'm getting 6.32 tokens/second. You: You are Hunter S. Thompson, write a 450 words essay on the LHC project. The writing must be in Thompson's style AI: Well, buckle up, kiddos, because we're about to dive into the twisted world of particle physics and the Large Hadron Collider, or as I like to call it, the Doomsday Machine. This beast of a thing is buried deep beneath the Swiss countryside, a 17-mile-long ring of death and destruction that makes the atom bomb look like a goddamn firecracker. Now, let me tell you, when the Swiss government first proposed this monstrosity, I thought they'd gone completely batshit crazy. But then again, this is the land of Toblerone and Heidi, so maybe I shouldn't be too surprised. The LHC is essentially a giant, underground racetrack for subatomic particles, where they're accelerated to nearly the speed of light before being smashed into each other like a couple of drunken frat boys at a kegger. The goal of this madness? To recreate the conditions that existed just after the Big Bang, when the universe was nothing more than a seething cauldron of energy and chaos. And if that doesn't sound like a recipe for disaster, I don't know what does. But let's not forget the real reason we're all here: the Higgs boson, or as it's more commonly…
AMD Ryzen 9: Likes itNVIDIA GeForce RTX: Likes it
+1. it's great that it compiles code faster, but i would also like when i'm writing code and hit enter, a newline is added and all lines under it goes down by 1 without any "lag". That's why i decided to go with a 5950x + 3090 rtx + gamer ram for my current "workstation" instead of a 5995wx + rtx 6000 + ecc ram for my current machine. even tho its not workstation rated, it seemed like the better way to go for quality of life feel. When the next generation of cpu/gpu/ram drops in the next couple eyars, i can re-evaluate the gamer vs professional offering to see if it matters.
AMD Ryzen 9: Likes itNVIDIA GeForce RTX: Likes it
> People who are into BSD don’t use BSD. This seems to be a reason for lacking hardware support. Ryzen 9 5650x - no issues AMD 5700 XT - no issues Nvidia 3070 - no issues Presonus - no issues While waiting for replacement morherboard I did a complete swap from AMD to Intel 4770k and no issues. Linux kernel panicked and Windows just BSOD. What issues are people having? Because the desktop runs like a charm, sound works. Internet and WiFi.
AMD Ryzen 9: Likes itNVIDIA GeForce RTX: Likes it
$52k gets you: - A 2019-era 28-core/56-thread Xeon W that considerably outbenches a contemporary Ryzen Threadripper but isn't anywhere close to a current one. - Two linked Radeon Pro W6800X Duos, with each Duo having 2x32GB GDDR6 GPUs - A $2,000 Mac Pro-exclusive, ProRes-decoding-only FPGA accelerator card to simultaneously playback six 8k/30fps ProRes RAW video streams, or 23 RAW or 16 422 streams at 4K, in FCPX and a few other video editing and transcoding apps - 1.5TB of DDR4 ECC RAM - 8TB SSD - Wheels instead of feet Some of these things are almost redundant. The Afterburner accelerator is a better value on a lower-end Mac Pro because its benefits are narrow and often on CPU-bound work. And the wheels are still a $400 option on a new Mac Pro. Your 7950X is a fantastic value at about 2/3 the raw performance of a 5995WX that costs considerably more. You probably don't have or need 1.5TB of RAM, but tasks dealing with multiterabyte-scale data sets (genomics!) can find a use for it. Your RTX 4080 is great for nearly all applications but short by about 8-64GB of RAM from the kinds of rendering or very large language model training that the target audience of a $52k Mac Pro might want to do with it.
AMD Ryzen 9: NeutralNVIDIA GeForce RTX: Neutral
Counting my monitors and the small network switch on my desk, I can hit around 750-800 watts sustained with a 5950X and a 3090 running compute intensive workloads as measured at the wall by a power meter. So yeah, there's plenty of head space but I have also hit circuit breaker issues recently on a proper (non consumer) workstation so the limit is fresh in my mind.