I'm aware of the data you're showing and I've never said otherwise; 16 MB of L3 cache is perfectly sufficient for now. I don't attach much importance to the PCIe 3.0 limit since flexibility is necessary if you have an iGPU.

Core 200 Raptor Lake Next w 2027 na starym LGA 1700. Trzeci refresh Raptor Lake (nie Bartlett Lake): desktop i mobile. Będzie współistnieć z nowszymi generacjami, wsparcie DDR4 i DDR5, bez większych nowości. @Jaykihn potwierdza: to dokładnie ten sam krzem Raptor Lake. Nowe modele desktopowe z włączonym iGPU: ➡️Core 7 8 12 65W ➡️Core 5 8 8 125W ➡️Core 5 6 4 65W z 24 MB L3 (nowa konfiguracja domyślnie 20 MB) ➡️Core 3 4 0 65W Brak Core i9, 8P 16E jest rozczarowujący. Intel wyciąga lekcję od AMD i przedłuża życie budżetowej platformy DDR4. Budżetowe buildy na płycie z 2022 roku jeszcze w 2027? #Intel #RaptorLake #CPU #LGA1700 #Core200

its an "ok" processor - but a cutdown version. as i said, as long as you use the iGPU its ok, but it often bottlenecks real GPUs due to its pcie 3.0 interface

i can never hate vulkan enough. who in their right mind made "descriptorBindingSampledImageUpdateAfterBind" optional even for vulkan 1.3. I am stuck with the 200-texture limit on an intel iGPU :/

no same shit a bit OC the igpu and that´s it 192gb ram that´s the only interesting bit vs 128gb

Raptor Lake Next Desktop iGPU Enabled SKUs C7 8 12 65W C5 8 8 125W C5 6 4 65W 24MB L3 C3 4 0 65W The i5 6 4 65W 24MB L3 is a new configuration (Default 20MB L3 for 6 4). The closest alternatives are the 13490f/14490f, which do not have integrated graphics enabled.

Benched RTS TTS on AMD Ryzen AI 9 HX 370 CPU vs. GPU, vulkan vs llama.cpp (very fresh) gist.github.com/eSlider/82dc… #amd #GPUs #igpu #bench

I mean..... you need a GPU for display out. So yes. You cannot operate a a PC / Mac without a GPU. Whether its an iGPU or DGPU. You need a GPU

AIもiGPUとNPUとわけるとそこそこ使えるので、タイパはいいですけど、めっちゃ速いってわけではないです

its Ryzen 7 with iGPU. currently I'm using Qwen 3.5 4B, SD1.5 FP16, SDXL Q8, and MioTTS for daily driver. I still have a list of models I am going to try with this hardware. you can check my benchmark on my localmaxxing profile localmaxxing.com/en/user/asa…

a little payer for our brother who fell for the wrong 5000Ryzen… except you wanted to go for the igpu for some reason - then its fine 😅

あれ、最近のアプライド通販さんは「超自由にグラボが選択出来るシステム」はやめちゃったのかな……選択を間違うとビカビカ光るだけで(iGPUなので)ゲームの出来ないデスクトップPCが届くアレ……🐺💦

That's insanely fast. I stupidly opened it at fullscreen on a 4k monitor with an old iGPU, still got 20 fps

My prototype was running 60 fps on older Intel iGPU. One million SDF volumes rendered. The source model for that volume was 1 million triangle mesh.

You can bypass the compositor and have a single process and display and get effective hard real-time out of it with a dedicated machine like we do for our IGs. Even for consumers, they can have a second GPU too. Many also do with an iGPU dGPU, although I’ve found that to be problematic to work with for general purpose users. In any case, the hardware is fine, it’s just the specific typical multitasking use cases managed by existing OSes and drivers that are the problem. I still believe GPUs offer better hard real-time guarantee potential than CPUs do on any of these OSes, especially when you’re writing to disk. NVMe drives are notoriously bad with their hardware IRQs randomly affecting any core, not just the one used to write to them. I’ve seen 3ms spikes from these. They cannot interrupt the GPU though.

That is the fun part i am pointing at My GPU is not listed but (2014) intel HD 4600 IS listed. My GPU have min 10x times performance of HD4600. Office laptops and convertible Tablets have only iGPU like HD4600. Steam Cherry picking what is listed. That survey is inaccurate

It could be missing hardware dependencies, the hardware could be barely acceptable to run these ports (GPU). Each project requires a minimum of a Nvidia Geforce 600 series card to run, which some Envy models may forgo for the iGPU. Do you have a GPU of that class or higher?

Microcenter sold the Msi prestige 14 358h only for a very brief period of time, and in fact this model is pretty much the only remotely affordable Panther Lake device out there that has the B390 GPU and 32 Gigs of RAM. Had a daily cron job doing a search for it and was able to land an open box unit at my local Microcenter. Will finally get to see what the much hyped igpu on Panther Lake is like when using Vulkan.