HNHacker News
TopNewBestAskShowJobs

kvemkon

391 karma · joined August 11, 2023

submissionscomments
kvemkon··on Mirror-bridge – Making Python bindings frictionless
> Before P2996, generating Python bindings required either:

I'm missing nanobind here.

> Method lookup. Python sees a.dot and searches for the dot attribute. It checks a.__dict__,

> Each attribute access (self.x) involves:

    Dictionary lookup in self.__dict__
Since C++ classes are fixed, I'd expect attributes to be declared once in python (slots).

One thing yet important for cross-language projects: callbacks (events). Would be nice if the sample could be extended.

kvemkon··on Micron Announces Exit from Crucial Consumer Business
> KLEVV

Just looked at standard desktop: still no 64GB 5600MT/s modules. CUDIMMs are missing 32GB.

> And don't forget about Nanya

BTW, what is the status of Elpida now?

kvemkon··on 1GB Raspberry Pi 5, and memory-driven price rises
Using digital chips instead of some novel analog approach is even greater waste.

> China's AI Analog Chip Claimed to Be 3000X Faster Than Nvidia's A100 GPU (04.11.2023)

https://news.ycombinator.com/item?id=38144619

> Q.ANT’s photonic chips – which compute using light instead of electricity – promise to deliver a 30-fold increase in energy efficiency and a 50-fold boost in computing speed, offering transformative potential for AI-driven data centers and HPC environments. (24.02.2025)

https://qant.com/press-releases/q-ant-and-ims-chips-launch-p...

kvemkon··on 1GB Raspberry Pi 5, and memory-driven price rises
> enterprise-level RAM sticks will operate with 9 8-bit-wide chips

Since DDR5 has 2 independent subchannels, 2 additional chips are needed.

kvemkon··on AI just proved Erdos Problem #124
> What we need is automated theorem discovery.

I've been thinking mathematicians have fun doing math, making discoveries, crafting proofs.

Does Tour de France & Co. make no sense since small, lightweight and powerful e-bicycles appeared?

Using computer as a helper like bicycles is one thing, using LLMs seems more like e-bicycle and is something another.

kvemkon··on DRAM prices are spiking, but I don't trust the industry's why
> DDR4 is EOL

11 years after specification released.

DDR4 makes it possible since 5 and half years to build 128GB AMD PC using 4x 32GB DDR4-3200 1.2V JEDEC modules. Only since half a year it is at last possible to build AMD PC with 128GB DDR5-5600 RAM. Because Ryzen DDR5 controller cannot operate with 2 sticks sharing the same channel at JEDEC speed/voltage.

DDR4 just cannot be EOL already because of unbuffered ECC! Even today it is not possible to have 128GB with 2 DDR5 sticks on AMD. Only 96GB.

And actually I'd expect 256GB (DDR5-8800 1.1V JEDEC) to be possible with DDR5. We're now only from 4000 to 5600 5(!) years since the specification was published. What if AMD achieves this only few months before DDR5 EOL?..

> DDR6 will be out in 2026 with consumer availability in 2027

... or not at all?..

kvemkon··on Is DWPD Still a Useful SSD Spec?
Now that PCIe 5.0 SSDs are available since 6+ months and you could backup your SSD with 15 GB/s but:

> you’re still writing at 1.5GB/sec.

Except of few seconds at the start, the whole process lasts as if you had PCIe 2.0 (15+ years ago). Having so fast SSDs there is no chance to make a quick backup/restore. And during restore you're second time in a row too slow.

It's crazy that instead of using slow PLC at the time of slow PCIe 1.0, back then fast SLC was in use. Now with PCIe 5.0 when you really need fast SLC, you get slow TLC or very slow QLC or even worse PLC coming.

kvemkon··on Unpowered SSDs slowly lose data
Does the refresh procedure take place at all?
kvemkon··on Unpowered SSDs slowly lose data
NVMe protocol introduced namespaces. Is it not the feature perfect for users to decide themselves, how to create 2 virtual SSDs with TLC and pseudo-SLC-mode, choosing how much space to sacrifice for pSLC?
kvemkon··on What Killed Perl?
> Python 2 to 3 took a really long time

It took AdaCore so long to port the plugin system of the GNAT Studio (GPS) to Python 3 (which seems to be a fraction of the whole code base), that even conservative Debian had to remove the whole GNAT-GPS package.

[1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1082332

kvemkon··on FPGA Based IBM-PC-XT
Dupe. Posted few days ago by author: https://news.ycombinator.com/item?id=45903083.
kvemkon··on COBOL to Kotlin via Formal Models (IR and Alloy and Golden Master)
> formal, verifiable modernization

Would it be possible to do the same to modernize a Kotlin program becoming legacy in the future to something even more modern?

kvemkon··on FFmpeg dealing with a security researcher
> FFmpeg would present a warning

Reminds me of gstreamer plugins being separated in "base", "good", "bad" and "ugly" sets.

kvemkon··on Apple loses UK App Store monopoly case, penalty might near $2B
> Why?

Just because of statistic.

But actually, of course, it would be fair to make individual prices. E.g. there is a photo-camera for 7000, it makes very very good photos. Someone makes every day money with photos and buys it. Someone else wants also very good photos for non-professional private use, family and even for posts in WEB (without a fee, only donations). But this someone else must pay the same 7000.

Needless to say that in case of individual prices, additional rules and locks will apply.

BTW, some electronic devices are sold cheaper for academics (universities, students).

kvemkon··on Apple loses UK App Store monopoly case, penalty might near $2B
E.g. EU companies could replace China 5G. But with US chips. While some new tiny FPGAs appeared recently (using obsolete manufacturing process), huge FPGAs from Xilinx(AMD)/Altera(Intel) cannot be replaced. China can make both US and EU calm down with rare earth materials. US can tame EU with chips and whatever else. And EU... no idea.
kvemkon··on Apple loses UK App Store monopoly case, penalty might near $2B
> Why are you excluding phone-like devices?

Actually not, for me *smart*phones are PC-like devices. Smartphone, especially modern one, is obviously a miniature personal computer.

> It's discriminating against the poor people of any country that also happens to have rich people in it.

I think, poor people are discriminated in every country generally if they need to buy products designed by an international company. While I've mentioned average income, it should be strictly speaking median income.

> Price discrimination

I mean that countries with lower median income deserves cheaper products. So that relatively there should be only minimal difference for poor people to be able to buy such products, regardless of the country of origin.

> International companies aren't "based" anywhere in particular except for on paper,

That's a good question. This "on paper" is crucial for all the legal stuff. US-based companies must obey not only EU rules but also US. Especially problematic is when they'd need to break a EU law to comply with the US law but they are never allowed to do this vice versa. This way I'd define where a company is based.

> anywhere is that the government says

It will be problematic due to multiple governments would be involved (typically US, EU, UK, China).

kvemkon··on Apple loses UK App Store monopoly case, penalty might near $2B
> and operating systems

This must be an essential requirement for all PC-like devices.

> regions

People in different regions have very different average income. As long as manufacturer scales prices accordingly, I think locking to this region might be appropriate.

> buy the unlocked ones there and use them anywhere

In the case if the price is already the worldwide highest one, you're right, no locking to the region is needed.

> EU could actually break them up

I think, EU could only achieve that a non-EU-based company leaves EU.

kvemkon··on Are hard drives getting better?
> replacement HDD that I could use for another 5 years

But the warranty lasts only 5 years since the purchase of the drive, doesn't it?

kvemkon··on Testing two 18 TB white label SATA hard drives from datablocks.dev
> drive has like millions of hours left of it's life.

It doesn't apply for the single drive, only for a large number of drives. E.g. if you have 100000 drives (2.4 million hours MTTF) in a server building with the required environmental conditions and maximum workload, be prepared to replace a drive once a day in average.

kvemkon··on How AWS S3 serves 1 petabyte per second on top of slow HDDs
> Full seek on a modern drive is a lot closer to 25ms than 8ms

Is "full seek" a synonymous for worst case time to reach a position occurring less than 1% of working time?

From the article: max seek time is 15.2 ms + additionally 0 to 8.3 ms of rotational latency.

Reordering of sector accesses by NCQ should reduce worst case scenario occurrences.

kvemkon··on Deep Dive into SATA, USB and PCI Express on AMD Turin
> only 16MB flash

Wow, that's nice catch! Since even desktop mainboards have 32MB it is something that wouldn't attract attention.

kvemkon··on Deep Dive into SATA, USB and PCI Express on AMD Turin
There is a rather cheap mainboard ASUS K14PA-U12 for 9004 EPYCs [1]. Strange, but it is missing support for 9005 Turin. How can this be (except of 400W+ models)? What do you think, could there be some true hardware issue making it impossible to workaround it in BIOS and thus no chance for coreboot either?

[1] https://servers.asus.com/products/servers/server-motherboard...

kvemkon··on CPU Utilization is Wrong (2017)
Related recent discussion: "%CPU utilization is a lie".

https://news.ycombinator.com/item?id=45110688

kvemkon··on GigaByte CXL memory expansion card with up to 512GB DRAM
I'd have rather a question why we had single (or already dual) core CPUs with dual-channel memory controller and now we have 16-core CPUs but still with only dual-channel RAM.
kvemkon··on GigaByte CXL memory expansion card with up to 512GB DRAM
Micron DDR5-5600 for 900 Euro (without VAT, business).
kvemkon··on 996 Working Hour System
Main discussion now: https://news.ycombinator.com/item?id=45149049
kvemkon··on Like Intel before it, AMD blames motherboard makers for burnt-out CPUs
Isn't it the small embedded support CPU core within the CPU I/O die that controls voltages and frequencies? Sure it depends of what is configured in BIOS.
kvemkon··on GMP damaging Zen 5 CPUs?
And the close-up photos of the socket with pins are missing.
kvemkon··on SSD-IQ: Uncovering the Hidden Side of SSD Performance [pdf]
> Vendors downplay the idiosyncrasies of specific SSD models by marketing their devices using four “headline” throughput metrics: sequential read, sequential write, random read, and random write.

For SOHO yes, where no serious database usage is expected. But server/datacenter SSDs are categorized: read-intensive, write-intensive and mixed-usage.

kvemkon··on Faster Index I/O with NVMe SSDs
> 128 KB appears a point of diminishing returns, larger block sizes yield similar or worse performance.

Indeed, 128 KB is a well-known long lasted optimal buffer size [1], [2].

Until it has been increased to 256 KB recently (07.04.2024) [3].

[1] https://github.com/MidnightCommander/mc/commit/e7c01c7781dcd...

[2] https://github.com/MidnightCommander/mc/issues/2193

[3] https://github.com/MidnightCommander/mc/commit/933b111a5dc7d...

← PreviousPage 4 of 10Next →