4 ms·
Is that a bottleneck anymore? Maybe something that was appreciable in 1994 isn't a big performance hit in 2026?
by floxy 3mo ago
Is that a bottleneck anymore? Maybe something that was appreciable in 1994 isn't a big performance hit in 2026?
- Analemma_ 3mo agoIsn't it even worse now? The faster computers get, the more that software-induced bottlenecks hurt.
- Manozco 3mo agoI'm not a Hurd contributor ( or any OS kernel / library ). I'm not sure this holds today. One example I can see is related to crypto. We used to have specific hardware for computing cryptography functions but it's now handled directly in standard hardware and the software has not evolved ( but it's been faster and faster to compute checksum functions )
- monocasa 3mo agoI mean, every major arch has crypto specific acceleration instructions.
- doubled112 3mo agoI could see just powering through because modern hardware is fast enough to do so being an option. Adding 20% to a 10 second operation is a lot longer on a wall clock than adding 20% to a 1 second operation.
- monocasa 3mo agoIf anything, Mach style IPC has gotten worse because of the greater overhead if kernel/user transitions. Spectre really did a number on context switching overhead.
- 0cf8612b2e1e 3mo agoWhat even is a 20% delta in performance? Like a five year old difference in hardware design. If I could get a micro kernel with a better security profile than Linux, that doesn’t sound so bad. I could easily be using a ten year old chip for my day to day browsing and probably wouldn’t notice.
- yjftsjthsd-h 3mo ago> If I could get a micro kernel with a better security profile than Linux, that doesn’t sound so bad. I suggest that you look at genode/sculpt:)
- Analemma_ 3mo agoPeople lost their shit when Spectre mitigations added 20% overhead to kernel calls though. Why wouldn't they do the same for microkernel overhead?
- floxy 3mo agoIf people really cared about performance, we'd be doing something other than the bloated way we have been doing things. https://winworldpc.com/product/qnx/144mb-demo https://winworldpc.com/product/qnx/144mb-demo
- pjmlp 3mo agoThe same people are perfectly fine with Electron crap all over the place, they don't really care about overhead.
- lanstin 3mo agoNot true in a data center - 20% less work for same CPU is like 25% more cores needed for any applications where the scaling is driven by CPU utilization. That is $ - and people have made things so far from efficient many medium size companies have 100k cores to pay for.
- floxy 3mo agoI'm having a hard time believing that data centers are chomping at the bit to use the Hurd.
- floxy 3mo agoHurd is like 36 years old at this point. More like adding 20% to a 10 second operation, compared to adding 20% to a 10ms operation.
- naasking 3mo agoYes, it's worse because processors clocks are much faster than memory clocks.
- panick21_ 3mo agoYes, you want to do more IPC, its basically linear with the increase in performance. So the amount of overhead stays the same basically. At least in IPC benchmarks and those tend to be important. Making that fast is fundamental to any OS.
- floxy 3mo agoSo for an I/O heavy program like a web browser, what percentage of CPU time is spent in syscalls? Someone must have that data for Firefox. 1%? 10%? And then that moves to 1.1% (or 11%)? And then every other program is even less affected? Curious if anyone is in the know.
- panick21_ 3mo agoIt depends on the OS, some OS have some way to do IO that gets around that. iouring or things like that. I don't know how Hurd does IO. Most microkernals have methods to do IO in a different way. If they don't then likely they would pretty slow when doing I/O heavy operations.
- lanstin 3mo agoFirefox isn’t io bound it is spending a ton of CPU parsing network buffers and rendering UI. Nginx is IO bound in general - for my headless go stateful server, i could get syscall6 up to 20% of the profile time. It was a fight between GC in the gRPC code allocating for headers and the syscall6 for top single profile footprint.
- floxy 3mo agoLooking forward to reading the results of your benchmarking Hurd vs. Linux.