Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Why PS4 downloads are so slow

snellman.net

Post reply on HN
Print view View original

31–40 of 201 posts

Re: Why PS4 downloads are so slow

#31

Post by jsnell »

foobarbazetc wrote: post #29

The CDN thing is an issue too. Using a local DNS resolver instead of Google DNS helped my PS4 speeds. The other "trick" if a download is getting slow is to run the in built "network test". This seems to reset all the windows back even if other things are running.

> The other "trick" if a download is getting slow is to run the in built "network test". This seems to reset all the windows back even if other things are running.

I did not see that in my testing. The built in speedtest runs with a large receive window, but the store downloads are not affected. (You can see an example in the last graph; there's a speedtest early on that has a receive window 100x larger than the PSN downloads that are crawling along).

It's probably just a placebo.

Re: Why PS4 downloads are so slow

#32

Post by amluto »

ploxiln wrote: post #18

Reminds me of how Windows Vista's "Multimedia Class Scheduler Service" would put a low cap on network throughput if any sound was playing: http://www.zdnet.com/article/follow-up-playing-music-severel... Mark Russinovich justified it by explaining that the network interrupt routine was just too expensive to be able to guarantee no glitches in media playback, so it was limited to 10 packets per millisecond when any med…

I thought Windows had a fancy interrupt priority system that should, in principle, allow sound playback to preempt network interrupts? AMD64 added a fancy feature (CR8 access to the task priority register) just to accelerate interrupt prioritization.

(This is all very vague memory. I know how this stuff works on Linux. Linux does not have interrupt priorities.)

Re: Why PS4 downloads are so slow

#33

Post by radicalbyte »

ShabbosGoy wrote: post #27
ploxiln wrote: post #18

Reminds me of how Windows Vista's "Multimedia Class Scheduler Service" would put a low cap on network throughput if any sound was playing: http://www.zdnet.com/article/follow-up-playing-music-severel... Mark Russinovich justified it by explaining that the network interrupt routine was just too expensive to be able to guarantee no glitches in media playback, so it was limited to 10 packets per millisecond when any med…

Thanks for the links and reading material. It's amazing how important the scheduler is in an OS. I'm no expert on OS-level programming, but couldn't they have just used a simple round robin scheduler?

Just look at the process you have running, 90% of them are doing nothing. Round Robin is a waste for an OS.

Re: Why PS4 downloads are so slow

#34

Post by apenwarr »

MBCook wrote: post #16
tortasaur wrote: post #14

Earlier quoted context omitted.

I want that, as it normalizes the need for large upload bandwidth. The only reason it's such a scarce commodity now is that households aren't perceived as needing that bandwidth. The proliferation of legal peer-to-peer systems will hopefully drive the spread of symmetric connections.

One of the reasons I switched off of AT&T uVerse is they seemed to give you the absolute minimum upload bandwidth needed to achieve the advertised download bandwidth. So if you were downloading or streaming and uploaded anything more than a small image your download speed would plummet as you ran out of upstream bandwidth. Online backups/syncing? Hope you're not trying to do anything else that requires bandwidth.

That's not quite right. The most common problem with (especially, but not only) DSL is bufferbloat, where you fill up your upload buffers, and this greatly increases RTT. Then your ACKs in response to downloaded content will take a long time to get through the queue and tell the server it's okay to send more data. You would have to have a truly tiny upload speed limit (like 0.01 Mbps) for the actual ACK throughput to be the limiting factor.

You can work around bufferbloat if you have a DSL modem with fq_codel (rare), or if you insert a router with a rate limiter and fq_codel, so basically you never fill up your DSL modem's queue.

Re: Why PS4 downloads are so slow

#35

Post by PaulHoule »

Sukera wrote: post #24
PaulHoule wrote: post #17

Earlier quoted context omitted.

It is astonishing how oblivious people who are "paid to spend time on the product" can be about the products they ship. Case in point, Microsoft shipped a EXPLORER.EXE with Windows 10 that is 20% whitespace because of an Adobe invented metadata format that I bet nobody involved with Windows ever thought about it. See http://ontology2.com/essays/LookingForMetadataInAllTheWrongP...

> Quick note: The situation is not different on Linux, MacOS, or other operating systems because all executable formats have some way to embed images. What publishers should do is remove unncessary metadata before publishing, which is easy to do in the case of the PNG because we can simply omit the iTXt chunk. From your link - so I guess the situation on linux isn't different? Or am I missing something, other than th…

I don't expect it to be different on Linux or any other operating system.

Re: Why PS4 downloads are so slow

#36

Post by g09980 »

Want to see something like this for (Apple's) App Store. Downloads are fast, but the App Store experience itself is so, so slow. Takes maybe five seconds to load search results or reviews even on a wi-fi connection.

Re: Why PS4 downloads are so slow

#37

Post by peterwwillis »

PaulHoule wrote: post #17
erikrothoff wrote: post #13

Totally unrelated but: Dang it must be awesome to have a service that people dissect at this level. This analysis is more in depth and knowledgable than anything I've ever seen while employed at large companies, where people are literally paid to spend time on the product.

It is astonishing how oblivious people who are "paid to spend time on the product" can be about the products they ship. Case in point, Microsoft shipped a EXPLORER.EXE with Windows 10 that is 20% whitespace because of an Adobe invented metadata format that I bet nobody involved with Windows ever thought about it. See http://ontology2.com/essays/LookingForMetadataInAllTheWrongP...

Wasting space is a byproduct of having too much space. Just because it's bloated does not mean they didn't know it was bloated. They probably didn't care.

As time goes by, minimum hardware requirements for operating systems increase. This isn't because it is more difficult to write software, or the hardware stopped being capable of doing the same thing it used to do. It's because software design wastes more resources as those resources become increasingly available.

Today, people ship statically compiled binaries of tens to hundreds of megabytes in size because it is convenient to do so. That would have been absolutely insane before increased bandwidth and storage space made it practical. I have an internet connection so fast my wifi card can't use all the bandwidth.

What a time to be alive.

Re: Why PS4 downloads are so slow

#38

Post by jwilk »

Sukera wrote: post #24
PaulHoule wrote: post #17

Earlier quoted context omitted.

It is astonishing how oblivious people who are "paid to spend time on the product" can be about the products they ship. Case in point, Microsoft shipped a EXPLORER.EXE with Windows 10 that is 20% whitespace because of an Adobe invented metadata format that I bet nobody involved with Windows ever thought about it. See http://ontology2.com/essays/LookingForMetadataInAllTheWrongP...

> Quick note: The situation is not different on Linux, MacOS, or other operating systems because all executable formats have some way to embed images. What publishers should do is remove unncessary metadata before publishing, which is easy to do in the case of the PNG because we can simply omit the iTXt chunk. From your link - so I guess the situation on linux isn't different? Or am I missing something, other than th…

I'm not quite sure what exactly the author's point is.

Yes, you could embed images in ELFs, too, if you really wanted, but why? There's no standardized way to do it, and putting the images in separate files is easier.

Re: Why PS4 downloads are so slow

#39

Post by dfee »

g09980 wrote: post #36

Want to see something like this for (Apple's) App Store. Downloads are fast, but the App Store experience itself is so, so slow. Takes maybe five seconds to load search results or reviews even on a wi-fi connection.

I’m glad I’m not the only one who’s noticed this.

Re: Why PS4 downloads are so slow

#40

Post by sharkbot »

amluto wrote: post #32
ploxiln wrote: post #18

Reminds me of how Windows Vista's "Multimedia Class Scheduler Service" would put a low cap on network throughput if any sound was playing: http://www.zdnet.com/article/follow-up-playing-music-severel... Mark Russinovich justified it by explaining that the network interrupt routine was just too expensive to be able to guarantee no glitches in media playback, so it was limited to 10 packets per millisecond when any med…

I thought Windows had a fancy interrupt priority system that should, in principle, allow sound playback to preempt network interrupts? AMD64 added a fancy feature (CR8 access to the task priority register) just to accelerate interrupt prioritization. (This is all very vague memory. I know how this stuff works on Linux. Linux does not have interrupt priorities.)

I could be wrong, but my understanding of some sound cards is that they have essentially a single memory buffer that they are reading from when instructed to play a sound. Most sound cards let the OS split the buffer into two halves and raise an interrupt when one half of the buffer completes playing.

Interrupt prioritization doesn't help much, because the sound data is likely being generated from user mode, and the playback complete interrupt is being handled in kernel mode, which would require a transition back for further processing. When receiving data from a network card, no transfer back to user mode is required and will have implicit priority over sound generation. Therefore, network processing is likely to starve out sound generation barring choices like those described.

(This is somewhat hearsay, corrections welcome :))

Post reply on HN