Zum Hauptinhalt springen
Android Tipps & SecuritySamsung zieht am 31. März 2027 den Stecker(23.10.2026 um 06:30 Uhr)
•
Android Tipps & SecurityID Lupo der ID Up: VW will 2027 die Preismarke von 20.000 Euro knacken(23.10.2026 um 06:00 Uhr)
•
Sicherheitslücken (CVE)PHP2Shell exposes the weakness in coordinated vulnerability disclosure(23.10.2026 um 02:39 Uhr)
••••
AI & KI NachrichtenQ&A: How AI came to model the “messy” human mind(23.10.2026 um 06:00 Uhr)
••••
Android Tipps & SecuritySamsung zieht am 31. März 2027 den Stecker(23.10.2026 um 06:30 Uhr)
•
Android Tipps & SecurityID Lupo der ID Up: VW will 2027 die Preismarke von 20.000 Euro knacken(23.10.2026 um 06:00 Uhr)
•
Sicherheitslücken (CVE)PHP2Shell exposes the weakness in coordinated vulnerability disclosure(23.10.2026 um 02:39 Uhr)
••••
AI & KI NachrichtenQ&A: How AI came to model the “messy” human mind(23.10.2026 um 06:00 Uhr)
••••
Intelligence View
⚡ tsecurity.de Intelligence

Tokens per second is a memory bandwidth number. Measure it in 2 seconds.

A 40B model that does 125 tok/s on a laptop, and a 7B model that does 1.5 tok/s on a 2011 CPU with the same amount of RAM. Same ballpark of memory, ~80× …

Beitrag
0
Seite
0
↗ Quelle (DEV Community)
Social ReaktionenReagiere als Erste:r — dein Feedback zählt!

A 40B model that does 125 tok/s on a laptop, and a 7B model that does 1.5 tok/s on a 2011 CPU with the same amount of RAM. Same ballpark of memory, ~80× difference in speed. Almost every explanation you will read for that gap is wrong, and the wrong explanation costs people money. Here is the number that actually decides it: how many bytes your... Weiterlesen: Tokens per second is a memory bandwidth number. Measure it in 2 seconds.

Intelligence Digest — free The top security news daily · unsubscribe anytime
Community Rating & Social Proof
⭐ Leser-Wertung
–
Ø / 5.0
No reader ratings yet
Was this page helpful?
0
Jetzt Bewertung abgeben
🟢 Niedrig (1-3.9) 🟡 Mittel (4.0-6.9) 🔴 Hoch (7.0-8.9) Kritisch (9.0-10.0)

Related story clusters & sources (vector AI)

Pull to refresh
Nächster Beitrag