Re: AVX-512 for LAME: code to try, if you have the hardware

Alexander Leidinger via Lame-dev <[email protected]> Tue, 28 Jul 2026 11:58:16 +0200
Newsgroups gmane.comp.audio.mp3.lame
Organization No organization, this is a private message.
Message-ID <[email protected]>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