zadikian

8 ms·

181.
by zadikian 8mo ago
Common one is if you want two structs deemed "equivalent" based on a few fields to get the same ID, and you're only concerned about accidental collision. There are valid use cases for that, but I've also seen it misused
182.
by zadikian 8mo ago
Yeah v4 is the goto, and you only use something else if you have a very specific reason like needing rough ordering
183.
by zadikian 8mo ago
Mainly the first thing. If your struct is already serial, of course serialization will be easy.
184.
by zadikian 8mo ago
Serdes time can be significant. There are use cases for the zero copy formats even though they use more space. Likewise bit-packed asn1 is often slower than byte-aligned.
185.
by zadikian 8mo ago
Indeed Capnproto is more optimized for serdes time than space usage.
186.
by zadikian 8mo ago
Remember it's the first derivative. The title and chart suggested at first that there are fewer tech jobs vs the start, but really it must be way more. I still kinda want to see this going back to 2000. That must be the biggest tech cr
187.
by zadikian 8mo ago
Pretty sure protobuf used a header to track field presence within a message, similarly to what this article does. That does have its own overhead you could avoid if you knew all fields were present, but that's not the assumption it mak
188.
by zadikian 8mo ago
It's confusing because a few comments up is "for the vast majority of people single core performance is all they care about, it's also cheaper" which is unrelated to ECC.
189.
by zadikian 8mo ago
12% for the DIMMs only, but with Intel you need Xeon and its accompanying motherboard for it. Someone said AMD "kinda" lets you do ECC on consumer hardware, not sure what the caveats are besides just being unbuffered. Assuming tha
190.
by zadikian 8mo ago
In theory, do you need a single machine for any of that, or would it be cheaper to use a low-availability cloud cluster? Tests are totally independent, and builds probably parallel enough.
191.
by zadikian 8mo ago
I'm more concerned how the Mac filesystems don't have payload checksums.
192.
by zadikian 8mo ago
Yeah you're right. But now I'm wondering if it preferred writing to swap over evicting the cache at that time... can make sense if the cache is being hit more than the pages it's swapping out. 13GB app usage is also odd. I fe
193.
by zadikian 8mo ago
There were several years where used cheese grater Mac Pros could be bought and upgraded for very cheap, and were still not too outdated. I only replaced my MacPro4,1 when the M1 mini came out, mainly cause of wattage.
194.
by zadikian 8mo ago
Servers and pro workstations normally have ECC RAM.
195.
by zadikian 8mo ago
Right now with "green" memory pressure and not so many Safari+FF tabs open, I see 16GB physical, 13.41 used, 2.68 cached, 2.32 swap, 6.57 app, 2.52 wired, 3.89 compressed. Why is there swap used when I have free memory? I don&#x27
196.
by zadikian 8mo ago
Only if jailbroken with openssh installed, right? The old default "alpine" ssh login
197.
by zadikian 8mo ago
Pretty sure this is a marketing limitation more than a technical one.
198.
by zadikian 8mo ago
I've thought about this for flight delays because there are a lot of insiders (literally inside the plane) who would give you better info than what flight trackers show. And the airport screens are the least accurate, like showing a 20
199.
by zadikian 8mo ago
It's easier to make money doing unexpected things
200.
by zadikian 8mo ago
VMs or huge builds can burn through that fast. Say 3 simultaneous Android emulators, or building Android itself
201.
by zadikian 8mo ago
Yes, I had a base 2015 MBP with 8GB RAM until recently. It was fine for light local dev: Node or Python backends + Postgres + a small Linux VM. And personal stuff like email/browsing. Wasn't ok for heavy IDEs like Android Studio,
202.
by zadikian 8mo ago
They might understand e2ee but not care.
203.
by zadikian 8mo ago
Was going to say something like this. If you're good at keeping things simple, it will help you deliver impact which can get you promoted.
204.
by zadikian 8mo ago
Oh, that's irritating
205.
by zadikian 8mo ago
I've seen scenarios where HT doesn't help, iirc very CPU-heavy things without much waiting on memory access. Which makes sense because the vcores are sharing the ALU. Also have seen it disabled in academic settings where they want
206.
by zadikian 8mo ago
Guessing this is especially hard to automate with peripherals involved. I once had a workload slow severely because it was running on the NUMA node that didn't share memory with the NIC.
207.
by zadikian 8mo ago
Wait, does a single CPU chip have numa within it now, or are you only talking about multi-socket machines?
208.
by zadikian 8mo ago
The real improvement was the number of displays supported and, in some cases, removing the Touch Bar and adding HDMI/SD.
209.
by zadikian 8mo ago
Yeppp. Is there no way for Electron apps to use a shared engine if one is available?
210.
by zadikian 8mo ago
They could resell, but maybe another way to phrase this, tying the screen to the obsolete computer greatly reduces the useful lifespan of the screen. But at that time, DisplayPort didn't do enough bandwidth to have that kind of display

More ›