I play my games in 4K, and this GPU performs really well. 💪 If you're looking for a graphics card for 4K gaming, I highly recommend it based on my experience.
Last time I ever did anything physically to my PC was back in 2010....yup that long ...it sure was time for an upgrade and I figured I would start off with AsRock IntelArc Challenger B580....
Hopefully new parts won't give me any trouble for as long as the old ones :D
Wish me luck!
Hi i'm just wondering why tf is the GPU util on AC8 is sitting at 8-10% on all settings? what is wrong with the game or is it the drivers, anybody that's knowledgeable about PC have any answer for me? Thanks.
Intel Core Ultra 270K Plus
16GB Ram
Intel Arc B580
This whole project actually started with me just trying to get Strata working on my Intel Arc GPUs. I wasn't originally planning on making a fork or adding a bunch of features. I just wanted to get the engine running on my hardware.
But as I got things working and started benchmarking, I noticed something interesting. A single request was generating around 40–45 tokens/sec, while multiple concurrent requests could make use of a lot more of the server's processing capacity.
That got me thinking. Since I'm the only one using my server, why not try putting those extra batch slots to work on a single complicated request?
That's basically how Strata Void started. It's an experimental fork of Niko1221's Strata, with a feature I've been working on called Task-Parallel Requests.
The idea is pretty simple. Instead of having the model work through one big question sequentially, it can break suitable requests into smaller independent tasks, process them concurrently using the same loaded model, and then combine everything into one final response.
You can enable it with "task_parallel": "auto". Simple questions go through normally, while more complicated requests can be split up if the planner thinks it would help.
I ran 10 different complex tasks with a 32K context window, using cold runs with one test per task and mode.
The important distinction is that this doesn't actually increase single-stream token generation speed. That's still around 40–45 tok/s on my setup. What improves is the total time it takes to finish a complicated request.
For reference, my setup is:
All of the model's expert weights are loaded into VRAM, so there's no RAM spilling involved in these benchmarks.
I've also been experimenting with faster prompt processing on Intel Arc, shared-context caching between subtasks, and larger context configurations up to 128K.
This isn't some magical free performance boost. Splitting a request means additional planning, processing, and synthesis, so it uses more total compute. It's also not useful for every prompt, and I've seen cases where parallel tasks make mistakes or disagree on numerical results.
The benchmarks are preliminary and all come from one machine and model, so I'm definitely not claiming everyone will see the same improvements.
Also, credit where it's due: the underlying engine comes from Niko1221 and the Strata contributors. I worked on the direction of this fork and its evaluation, with Claude assisting heavily with implementation, debugging, and benchmarking. Task parallelism itself isn't a new concept; I wanted to see how well it could work integrated directly into Strata.
I'm sharing this because I'd genuinely love to see how it performs on other hardware, especially different Intel Arc configurations or NVIDIA CUDA setups.
If anyone feels like experimenting with it, I'd love to hear what works, what breaks, or whether the performance improvements hold up on your system. Contributions and bug reports are welcome too.
GitHub: https://github.com/Jumbomuffin777/Strata-Void
The repository has the setup instructions, benchmark methodology, and raw results if anyone wants to dig into the details.
Arc owners get the short end of the stick on benchmark coverage, far fewer per-game FPS numbers floating around than for Nvidia or AMD. So I wanted to share a free tool that covers Arc properly.
You enter your Arc card (B580, A770, A750, A380 and the rest) and your CPU, and it goes through your whole Steam library and shows, per game:
How the numbers work, so you know what you are looking at: they come from each game's requirements plus a performance model, and where I have real measured benchmark data the estimate is calibrated against it. Each number is labelled "Benchmark" (corrected against measured data) or "Estimate" (requirements-based, can be off), so you always know which is which.
Where I would really like Arc owners' help: Arc is the thinnest part of my benchmark data, so if a number looks off for your exact card and a game you know, tell me the card, the game and your real FPS. Owners can report measured FPS right on the site, and that feeds straight back into the calibration for Arc specifically.
I have an Ultra Core 5, 250k plus and Intel Arc A750 in my machine and was wondering if games have been falling back to the igpu?
I do have my DP plugged in from the monitor to gpu, but was still wondering because I noticed in FFXVI on my AIO reading, and in task manager, the cpu was above 30% while the gpu was at 1-5% at any given moment.
I disabled the Igpu in device manager to see if it improves my game performance/ if those numbers change, and was just curious if anyone else experienced something similar, or if it's common, as well as what's going on??