411–420 of 565 posts
Earlier quoted context omitted.
I'm not sure what you're talking about. Any app compiled using LLVM 17 (2023) can use SME directly and any app that uses Apple's Accelerate framework automatically takes advantage of SME since iOS 18/macOS 15 last year.
Most apps do not use Apple’s Accelerate framework.
Earlier quoted context omitted.
What do you base this on? I use it in all my products and I don't see why any performance-sensitive dev outfit wouldn't at least consider using it.
Most dev outfits are not performance-sensitive
Earlier quoted context omitted.
Nice. Lots of people still claim that M1 is super duper fast but now we're at almost twice the performance!
I claim that the M1 (macbook air) is fast. I also claim it's about half the speed of my similarly priced desktop of the same vintage for slow tasks that I care about. So I guess we've caught up with the desktop now. Actually I assume we caught up awhile ago if I used the beefy multi core MX-Ultra variants they released, really just the base model has caught up. On the other hand I could have spent four times as much…
On the move laptops will always be a bit slow because all the tricks to save idle usage don't help much when you're actually putting them to work.
Earlier quoted context omitted.
you could try moving your dev environment & docker into a linux VM.. could do something similar to this: https://medium.com/carvago-development/my-docker-on-macos-pa... I mention this because I'm guessing you might be using Desktop Docker which is kinda slow.
Or maybe just switch to OrbStack: https://orbstack.dev/
Earlier quoted context omitted.
I dunno, working with M1 daily I struggle with resource contention and slow py/js builds. I'd love something faster when work provides me with updated device.
My m1 16gb pro gets throttled (and manually restarted) every time vs code hits 100gb ram usage when running jest unit tests. Don’t know who to blame but 99% sure the problem is my source code and not the machine.
Earlier quoted context omitted.
The only valid benchmark for that is to use the application you intend to use as a benchmark. Even embarassingly parallel problems can have different characteristics depending on their use of memory and caches and the thermal characteristics of the CPU. Something that uses only L1 cache and registers will probably scale almost linearly in the number of cores, except for thermal influences. Something that uses L2, L3…
You're essentially just arguing that all general-purpose benchmarks are worthless because your application could be different. Suppose I run many different kinds of applications and am just looking for an overall score to provide a general idea of how two machines compare with one another. That's supposed to be the purpose of these benchmarks, isn't it? But this one seems to be unusually useless at distinguishing bet…
But in your example, deciding between 24 cores with somewhat higher frequency or 32 cores with somewhat lower frequency based on some general-purpose benchmark is essentially pointless. The difference will be small enough that only the real application benchmark can tell you what you need to know. A general purpose benchmark will be no better than a coin toss, because the exact workings of the benchmark, the weightings of it's components into a score and the exact hardware you are running on will have interactions that will determine the decision to a far greater amount. You are right that there could be shared or separate caches, shared or separate memory channels. The benchmark might exercise those, or it might not. It might heat certain parts of the die more than others. It might just be the epitome of embarassingly parallel benchmarks, BogoMIPS, which is a loop executing NOPs. The predictive value of the general purpose benchmark is nil in those cases. The variability from the benchmark maker's choices will always necessarily introduce a bias and therefore a measurement uncertainty. And what you are trying to measure is usually smaller than that uncertainty. Therefore: No better than a coin toss.
Nice. The iPads generally measure around ~8% slower than the MacBooks, I guess for cooling reasons. So we should see approximately a 4400 single core Geekbench score for the MacBook series. This is nice. Single thread MacBook progression on Geekbench: M1: 2350 M2: 2600 M3: 3100 M4: 3850 M5: 4400 (estimated) https://browser.geekbench.com/mac-benchmarks
Nice. Lots of people still claim that M1 is super duper fast but now we're at almost twice the performance!
Earlier quoted context omitted.
Here’s the multi core Geekbench progression: M1: 8350 M2: 9700 M3: 11650 M4: 14600 M5: 16650 (estimated) This is assuming an 8% uplift as mentioned. Also nice.
I wish we could get something other than Geekbench for these things, since Geekbench seems to be trash. For example, it has the Ryzen 7 7700X with a higher multi-core score than the Epyc 9534 even though they're both Zen4 and the latter has 8 times as many cores and is significantly faster on threaded workloads in real life.