AMD Ryzen 9 9900X3D im Test: Benchmarks nach Datenblatt und als „Ryzen 5 9600X3D“ https://www.computerbase.de/artikel/prozessoren/amd-ryzen-9-9900x3d-test.91949/ #AMD #Ryzen
Oh, the #AMD NPU driver already got upstreamed into #Linux 6.14? Fancy that, I thought I had to build myself a dkms module to get it running on OpenSuse. 
Now I'm curious how involved it will gonna be to make something like #Codestral run on it. The usual tools are probably not yet well optimized to utilize that Co-Processor OOTB.
Flight: #AIC172
ICAO code: #8005EC
Registration: VT-ANB
Type: BOEING 787-8 Dreamliner
Operator: Air India Limited
Direction: 67° (East-Northeast)
Speed: 288.0 kmh
Altitude: 10025 ft
Category: Heavy
Route:
: London - #LGW
: Ahmedabad - #AMD
: Autopilot engaged
Previously seen: 44 times
Track:
https://globe.adsb.fi/?icao=8005ec
Photo by: ZhangerAviation
Link: https://hooks.geekyco.de/ps/8005ec
Dynamic Register Allocation on AMD's RDNA 4 GPU Architecture
https://chipsandcheese.com/p/dynamic-register-allocation-on-amds
I'm quite disappointed by subgroupRotate or any other subgroup shuffle instructions. For both RDNA3+ & Ada Lovelace+, I thought that these instructions would be faster than e.g. going back and forth to LDS, but they look super slow.
And not a single GLSL code using such instructions on GitHub... while they have been around for many years.
Are subgroup instructions just not worth? Not used? Abandoned?
RDNA3+ ISA manual says it is using LDS hardware (but not mem), and so?
NVIDIA publica el codi font de PhysX i Flow sota la llicència BSD-3.
Obre la porta a fer servir la tecnologia en gràfiques AMD i arreglar els jocs que no funcionen amb les RTX50.
https://www.techpowerup.com/335117/nvidia-physx-and-flow-made-fully-open-source
NVIDIA Engineer Fixes Early Linux 6.15 Performance Regression Affecting AMD GPU Drivers
https://www.phoronix.com/news/Linux-6.15-Early-AMD-Regression
Discussions: https://discu.eu/q/https://www.phoronix.com/news/Linux-6.15-Early-AMD-Regression
Dynamic Register Allocation on AMD's RDNA 4 GPU Architecture
https://chipsandcheese.com/p/dynamic-register-allocation-on-amds
Dynamic Register Allocation on AMD's RDNA 4 GPU Architecture
https://chipsandcheese.com/p/dynamic-register-allocation-on-amds
Flight: #AIC171
Registration: VT-ANY
ICAO code: #80073F
Callsign: #AIRINDIA
Operator: Air India Limited
Type: BOEING 787-8
Country: 
From: #AMD to #LGW
Speed: 905 kmh
Altitude: 11582 m
Distance: 4.2 km
Angle ∆: 70.1°
Direction ->: WNW
Track:
https://tinyurl.com/29rghs2s
History:
https://www.radarbox.com/data/mode-s/80073F
https://www.flightradar24.com/data/aircraft/VT-ANY
Photos:
https://jetphotos.com/photo/keyword/VT-ANY
Seen: 8x
Looked at it again: yeah, the AMD GPUs definitely have at least one power saving bug that makes them draw 4x the power at idle, including the latest Radeon RX 9070 -.-
That’s so disappointing, and also that it seems to be an issue with AMD cards for many years!
show me one idiot who will pay that much 
A fan blade broke off my 7900 XTX mid gaming session, I was playing the Last of Us 2 for about an hour when i heard a slight rattle noise from computer then within about 60-90 seconds it turned into loud clank and grinding noise, and i found the blade at bottom of my case.
Flight Radar Mechelen 
/*
* msecs_to_jiffies interprets all values above 2^31-1 as infinite,
* but we consider them finite.
* This hack is wrong, but nobody is likely to notice.
*/
https://elixir.bootlin.com/linux/v6.13.7/source/drivers/gpu/drm/amd/amdkfd/kfd_events.c#L910-L914
Flight: #AIC172
ICAO code: #80073F
Registration: VT-ANY
Type: BOEING 787-8 Dreamliner
Operator: Air India Limited
Direction: 91° (East)
Speed: 293.1 kmh
Altitude: 12825 ft
Category: Heavy
Route:
: London - #LGW
: Ahmedabad - #AMD
: Autopilot engaged
Previously seen: 11 times
Track:
https://globe.adsb.fi/?icao=80073f
Photo by: GZ-T16
Link: https://hooks.geekyco.de/ps/80073f