menaerus

30 ms·

541.
by menaerus 1y ago
But you never serialize a byte by byte over the network. You encode, let's say, 1000 varints and then send them out on the wire. On the other end, you may or may not know how many varints there are to unpack but since you certainly hav
542.
by menaerus 1y ago
Zen3 decode is 4-wide + 8 uOp cache, and dispatch backend is 6 uOps wide. Theoretically, it shouldn't be possible to have IPC larger than 6.
543.
by menaerus 1y ago
> LEB128 is slow and there’s no way around that. Actually there is - you can exploit the data parallelism through SIMD to replace the logarithmic with near-constant complexity. Classic approach indeed is very slow and unfriendly to the C
544.
by menaerus 1y ago
Ok, I can see how I didn't manage to fully explain myself - by replying to your "collisions" argument I actually meant something more than that and while it seemed obvious in my head what I wanted to say, I can agree that my
545.
by menaerus 1y ago
You're incredibly rude and also clueless but I'm nonetheless giving you a benefit of a doubt to understand if there is something that I'm missing. But you're right, this is now way over the top for your head.
546.
by menaerus 1y ago
Actually Apple M design can hit ~100GB/s of MBW with a single core. Something that many other (or basically none?) CPUs of the same range couldn't.
547.
by menaerus 1y ago
Interesting design. 16 CCDs / 16 CCXs / 16 cores. 1 core per each CCD. 1 CCX per each CCD. With 512MB of L3 cache this CPU should be able to use ~all of its ~10 TB/s of L3 MBW out of the box. How much is it going to cost you
548.
by menaerus 1y ago
> Whether there is a hash collision is independent of the workload, In read-only workload, searching through the key-value space there can't be a collision, no?
549.
by menaerus 1y ago
I don't follow. If I have an ingestion-heavy workload then my hash-map will certainly be stressed in a completely different way than if I have a read-only workload. The data that PGO collects during the former will be different than th
550.
by menaerus 1y ago
What if your "application" is a server-side code hosting multitude of different types of workloads, e.g. databases? I was never really convinced by the PGO benefits for such open-ended and complex workloads.
551.
by menaerus 1y ago
Unless you missed it from another comment in this thread, the optimization you made is based on a wrong assumption and misreading the asm code: "je .L3" will skip all the costs you consider.
552.
by menaerus 1y ago
The issue was that they were hand-optimizing the SIMD for a workload that doesn't even exist in their case so they ended up seeing one big nothing once they profiled the code with a new optimization - a classic trap with 95% "opti
553.
by menaerus 1y ago
Ok, this then goes to say that your approach doesn't work without applying whatever fixes to the vanilla models. What I'm trying to understand is the approach itself. Why does it and how does it work?
554.
by menaerus 1y ago
Quite strong claims. I would be interested to read more about those. Do you have any references you can point out? A quick research on EGFR and TP53 suggested they are more related to colorectal cancer and not the lung cancer.
555.
by menaerus 1y ago
What does your approach with dynamics weights has to do with those bugs? All those bugs seem uncorrelated to the technique.
556.
by menaerus 1y ago
> No, on the fact that there are plenty of known variants which increase the chance of lung cancer quite dramatically - moreso than smoking. But didn't you just say that it is the genetics and not the radon and/or smoking? Can
557.
by menaerus 1y ago
> The key reason to use Unsloth quants is because of our deep involvement in fixing critical bugs across major models sounds convincing, eh ... /s On the less cynical note, approach does look interesting but I'd also like to un
558.
by menaerus 1y ago
And you base your conclusion on top of the sample size of what, N=10?
559.
by menaerus 1y ago
I never heard of the radon cancer impact before and it took me 10 minutes to find couple of representative studies done for my local region which in fact did find the positive correlation between the areas with elevated radon measurements a
560.
by menaerus 1y ago
Right, I was scratching my head exactly for that reason too. Even if the analysis was correct it would still be a solution for the problem that doesn't exist.
561.
by menaerus 1y ago
> I don't think that a system that underperforms the average neurotypical adult human in "most" cognitive tasks would constitute AGI What makes you say that it underperforms? I ask because evidence strongly suggests that
562.
by menaerus 1y ago
Chunking the codebase that you entirely own into packages is as if you're intentionally wanting to make your life miserable by imposing the same kind of volatility that you would otherwise find in the development process of building th
563.
by menaerus 1y ago
Fun presentation, thanks! 72min ingestion time for ~81TB of data is ~1TB/min or ~19GB/s. Distributed or single-node? Shards? I see 50 jobs are used for parallel ingestion, and I wonder how ~19GB/s was achieved since ingestion
564.
by menaerus 1y ago
Severe limitations? You sound like an a*hole tbf. The tool clearly provides the value, and it's a magnificent display how something like a simple LLM can be used to make your everyday life easier without compromising other minor stuff,
565.
by menaerus 1y ago
Wait, whole (english speaking) web content dataset size is ~50TB?
566.
by menaerus 1y ago
I don't think such liability is really exercised all around the world, perhaps in the States it is. But how do you prove that psychotherapist's advice is what caused you the (psychological) damage? Court?
567.
by menaerus 1y ago
It's pretty naive, it cat's from /proc/mem and extracts MemTotal and MemAvailable: https://github.com/surban/memstop/blob/master/memstop.c#L40-... . In other words, that's not how
568.
by menaerus 1y ago
Ok, I wasn't making a comment about "rightwing grifters" although this is also certainly amenable to discussion but I was rather making a comment on why would some people find it contrary to what you said in your first commen
569.
by menaerus 1y ago
It's fine with me if you think that the bottle caps were a worthy move, ok? We are different in the fact that I prefer to discuss objectively with arguments and not using the collective bias as you seem to impose on me without actually
570.
by menaerus 1y ago
I haven't looked into each EU country but it's difficult to imagine that a salary of X * K EUR gross will become 2X * K EUR gross just because the country has switched to EUR. It's delusional at best.

More ›