1/2 FWIW Scuttlebutt fr: weibo, Reddit, messg boards, murky websites, suggest Apple has GOOD REASON to be increasingly more tight-lipped/not publishing numbers for its latest silicon: is falling behind! — M5 will be an ITERATIVE evolution of M4; not ready in 2025 for TSMC’s SoIC-MH 3D packaging!
March 17, 2025 at 1:12 AM
1/2 FWIW Scuttlebutt fr: weibo, Reddit, messg boards, murky websites, suggest Apple has GOOD REASON to be increasingly more tight-lipped/not publishing numbers for its latest silicon: is falling behind! — M5 will be an ITERATIVE evolution of M4; not ready in 2025 for TSMC’s SoIC-MH 3D packaging!
Why on earth does Apple refer internally to the M4’s P/E CPU cores as “Avalanche/Blizzard,” which harkens all the way back to the A15⁉️
The M4 P/E cores have more in common architecturally with the A18’s “Sirius/Nova” (ARMv9.x.x) P/E cores than even the A17’s “Everest/Sawtooth” cores(?) 🤔
The M4 P/E cores have more in common architecturally with the A18’s “Sirius/Nova” (ARMv9.x.x) P/E cores than even the A17’s “Everest/Sawtooth” cores(?) 🤔
March 16, 2025 at 12:38 AM
Why on earth does Apple refer internally to the M4’s P/E CPU cores as “Avalanche/Blizzard,” which harkens all the way back to the A15⁉️
The M4 P/E cores have more in common architecturally with the A18’s “Sirius/Nova” (ARMv9.x.x) P/E cores than even the A17’s “Everest/Sawtooth” cores(?) 🤔
The M4 P/E cores have more in common architecturally with the A18’s “Sirius/Nova” (ARMv9.x.x) P/E cores than even the A17’s “Everest/Sawtooth” cores(?) 🤔
Apple’s M5 had better not be like a 13% performance improvemt over the M4 — it should be AT LEAST +25% (but preferably more like 40%).
Apple would do well not to lag in the LEAST in terms of things like bleeding edge HBM RAM & the like.
Today’s HPC is using the fastest HBM the industry will bear…
Apple would do well not to lag in the LEAST in terms of things like bleeding edge HBM RAM & the like.
Today’s HPC is using the fastest HBM the industry will bear…
March 15, 2025 at 5:08 AM
Apple’s M5 had better not be like a 13% performance improvemt over the M4 — it should be AT LEAST +25% (but preferably more like 40%).
Apple would do well not to lag in the LEAST in terms of things like bleeding edge HBM RAM & the like.
Today’s HPC is using the fastest HBM the industry will bear…
Apple would do well not to lag in the LEAST in terms of things like bleeding edge HBM RAM & the like.
Today’s HPC is using the fastest HBM the industry will bear…
The custom Apple silicon used for AI workloads in its proprietary data centers are NOT M-SERIES CHIPS and are built with a different design philosophy—including MORE ADVANCED NPUs than M-series designs.
Apple “Project ACDC” chips focus specifically on AI inferencing w/greater efficiency and speed.
Apple “Project ACDC” chips focus specifically on AI inferencing w/greater efficiency and speed.
November 20, 2024 at 8:12 PM
The custom Apple silicon used for AI workloads in its proprietary data centers are NOT M-SERIES CHIPS and are built with a different design philosophy—including MORE ADVANCED NPUs than M-series designs.
Apple “Project ACDC” chips focus specifically on AI inferencing w/greater efficiency and speed.
Apple “Project ACDC” chips focus specifically on AI inferencing w/greater efficiency and speed.
Reposted by R2DHue
The M4 MacBook Pro features a quantum dot display
The M4 MacBook Pro features a quantum dot display
That wasn’t on the spec sheet.
buff.ly
November 15, 2024 at 7:00 PM
The M4 MacBook Pro features a quantum dot display
Reposted by R2DHue
Blender benchmark highlights how powerful the M4 Max’s graphics truly are
Blender benchmark highlights how powerful the M4 Max’s graphics truly are
Apple released its new MacBook Pros with M4, M4 Pro, and M4 Max chipsets last week – and early CPU benchmarks (via Geekbench) painted a very promising picture for Apple’s highest end M4 Max chipset. Now we’re taking a look at some GPU benchmarks, and the…
9to5mac.com
November 17, 2024 at 8:36 PM
Blender benchmark highlights how powerful the M4 Max’s graphics truly are