Root Access

12:00 PM

Is all of this correct? Is this true? Because:

That would mean main memory is actually attached and mapped as I/O or not? The CPU cache is the actual memory that uses this easy systembus (address data control bus) concept. And when we talk about the simplest form of physical addressing what do we mean? (0x0FFF, 010FFh) Memory locations of course. But that's not the case, since at least the 90`s when the Front-side bus concept was introduced.

Its a YES or NO question

asking for an explanation in addition to the yes/no answer

Journeyman Geek

@Junaga only the person who downvoted really knows

Also, being migrated dosen't mean its on topic here, or people consider it to be

Burgi

did you know you can't even pay for a drink with SE rep?

Dog

@Junaga What Burgi said

Journeyman Geek

@Burgi IKR?

I think the big issue for me with the question is there's no real 'issue' there

Oliver Salzburg

@Burgi If you award me a bounty of 500 rep on one of my answers, I will buy you a glass of fresh water

Journeyman Geek

12:04 PM

its very much a computer architecture/curiocity question

rather than a real, practical issue

Burgi

hehehe

lunch time!

Dog

> On a modern computer system, the bus connecting the CPU to everything else is now named front side bus.

Ermm no, modern computers (basically anything less than ten years old) don't have front side busses

Journeyman Geek

I thought the FSB was essentially dead for some reason.

and we're moving towards tight point to point, in-package interconnects for many things

and no real north/south bridge as things go on

Junaga

@all as answered to one commenter

Dog

@JourneymanGeek Yes. The IOH on Intel is basically attached via PCIexpress now

On AMD it's been HT for years

Junaga

12:06 PM

@DanielB it only appears to be obsolete/out of date, because from a physical point of view the front-side bus connects the CPU to I/O (that is the DRAM controller always which is also a processor and responsible for DMA) that is called northbridge

Dog

Today there's no memory bus at all.

Junaga

(this is a name form a hardware point of view) with the new Intel 5 Series this component is actually build into the CPU itself, en.wikipedia.org/wiki/Intel_5_Series#Design_concept look at the images on the right. But it still is the front side bus which connects the cache used by different cores to the DRAM controlle

Dog

There's an internal data bus (remnants of back side maybe?)

Journeyman Geek

@Junaga @ notion only works if someone's here

Dog

@Junaga There is no northbridge, it's in the CPU

So much has changed since those old articles were written...

Journeyman Geek

12:08 PM

yeah

Junaga

It downs't matter if its inside or not if you view form software side

Journeyman Geek

meh

Dog

Northbridge? Part of CPU now. Graphics? Part of CPU now. I/O? Part of CPU now. Memory controller? Part of CPU now. PCI controller? Part of CPU now.

Journeyman Geek

Software might abstract away many things

@Dog essentially you have a soundcard there too

Dog

Memory has always been I/O mapped since DMA

Junaga

12:09 PM

*System software point of view

Journeyman Geek

but you're asking about architecture

Dog

However strictly speaking DMA does't "Access memory without involving the CPU" since the memory is directly attached to the CPU

Junaga

@Dog thanks^^ that's an actual answer

Dog

That said the graphics card showing in LSPCI is also part of the CPU, and has never been attached via PCI

Junaga

it actually isn't the thing that is directly attached to the cpu which we considre memory in theory is the cache

Dog

12:10 PM

I suspect the PCI "bus" is just an emulation/compatibility layer to allow software to treat it as any other PCI device rather than a special CPU-internal proprietary bus

Junaga

see question

The CPU cache is the actual memory that uses this easy systembus (address data control bus) concept. And when we talk about the simplest form of physical addressing what do we mean? (0x0FFF, 010FFh) Memory locations of course. But that's not the case, since at least the 90`s when the Front-side bus concept was introduced.

Dog

Heck PCIexpress primarily exists as a PCI emulation layer, to make writing drivers and software easier

Junaga

@Dog you mean a bus bridge?

a controller attached to the first bus that creates the second bus that can be of the same or a different bus type as the first bus

that’s how I/O works, like since ever

Dog

Err me stupid. I/O is memory mapped not the other way round

Where's bob when you need him. Bob knows this sort of shit and he's never wrong

Junaga

I have the feeling no one takes me serious just because I have no rep. lul

Journeyman Geek

12:15 PM

We don't take anything seriously

Junaga

xD

Journeyman Geek

If you ask a question like that you shouldn't expect an answer here

you would expect discussion and arguement

Junaga

But you know what that measn if I am right about this?

Journeyman Geek

See "discussion" not "answer" ;p

WHATEVER Dave

12:28 PM

@Junaga: Know the feeling.

Dog

I can't understand half what you say, so no, I don't know what you mean

allquicatic

@Junaga the "new" Intel 5 Series is from 2008, which is 8 years ago

you seem to be living in the year 2008 with almost everything you say

stop reading obsolete textbooks or websites

Junaga

@allquicatic If You like to say something or judge me read the question first at leastxd

Dog

^^ What fwog said

WHATEVER Dave

I hate reading

Dog

12:34 PM

WHATEVER, dave

WHATEVER Dave

Especially questions

Junaga

if intel series is already old?

whats the new version?

WHATEVER Dave

whaddup dawwwg

Junaga

5

Dog

@Junaga The 170 series

allquicatic

12:35 PM

List of Intel chipsets

This article provides a list of motherboard chipsets made by Intel, divided into three main categories: those that use the PCI bus for interconnection (the 4xx series), those that connect using specialized "hub links" (the 8xx series), and those that connect using PCI Express (the 9xx series). The chipsets are listed in chronological order. == Pre-chipset situation == Early IBM XT-compatible mainboards did not have a chipset yet, but relied instead on a collection of discrete TTL chips by Intel: the 8284 clock generator the 8288 bus controller the 8254 Programmable Interval Timer the 8255 parallel...

Junaga

I actually googled this, wow

@Dog

thats a list no example

the true is

intel 4 series

is the actual used model

it is old yes

but is it out of use ?!

Journeyman Geek

my 5 year old PC has a Z77. Now we're at 107

Junaga

fcking no?!

Dog

Oldest working machine I have is a 6 series, so yes.

I'll be less than 0.1% of 4 series machines are still in use

And it's probably like ATMs and nuclear submarines that get hardware updates every 20 years

allquicatic

@Junaga What does "is the actual used model" actually mean? By who? Where? What is the purpose of saying that? What is your point? Sure -- it is technically possible that you could still be using an obsolete computer; but what is your purpose in discussing or learning about them, when the entire architecture changes about every 5 years?

Journeyman Geek

12:38 PM

Oldest working machine I have is a C2D ;)

Dog

8 mins ago, by Dog

I can't understand half what you say, so no, I don't know what you mean

@JourneymanGeek Oldest working machine you have is your brain!

:-P

Journeyman Geek

@Dog whether it works is debatable

Dog

@JourneymanGeek Well I know mine doesn't, hence my oldest being a 6 series

allquicatic

@JourneymanGeek funny story: the C2D is actually using an older than 5 series chipset IIRC

Dog

I think I did find a server company still renting out Core 2 machines the other day

@allquicatic Indeed, it used the 9 series! :-P

Journeyman Geek

12:39 PM

(my dad uses it as his main/only machine)

Dog

Series 5 was first-gen Nehalem core i7

allquicatic

@Dog because version numbers make so much sense, right?

9 to 5 to ... 170

Journeyman Geek

OVER 9000

Dog

@JourneymanGeek I WOULD BUY THAT RIGHT NOW

7 hours ago, by Psycogeek

WTF

Journeyman Geek

@Dog to be honest, my reply to this would be mean and too painful, and involve laptops with 120hz OLEDs...

Dog

12:42 PM

@JourneymanGeek I would buy that right now too

Junaga

guys... just read the question before arguing xd superuser.com/questions/1139912/…

Dog

Sadly the 120Hz OLED I've been waiting on since January, is still nowhere to be seen

Journeyman Geek

@Junaga you clearly misunderstand this place

Its a chat room. If you want actual answers to a question... its not quite the right place

Junaga

I am not asking from a hardware point of view...(the question says this). So it doesn't matte if we are in theory talking about a 5series or 9 series okeay, its wyane is the current model has 10 transistors more

the point if about the forn side buse being conncted to the DRAM controller

ad again it doesn't matte if that is part of the niorthbridge or not if it is build into the copu or not

thats is a phisicall view on the system and how the dies are made

for system software

THIS doesn't matter

acctually I just asked why my questionwas downvoted then someone said its unclear and the deisscussion began

Journeyman Geek

@Junaga that happens here

Junaga

12:45 PM

Now I am wasting my own and everyones else time, well gone leave now

Dog

Blocked

Junaga

lol

Dog

First test: Chrome, second: Edge.

Still surprised how huge a difference the browser makes in the test :-/

Fucks sake why does Chrome suck so much

allquicatic

0

A: Does a x86 computer use the pci bus and pci protocols to access memory?

Most of your question contains concepts that are Outdated (Intel has been releasing platforms and CPUs where these things are not true since at least 2011) Incorrect (Main memory access is not a form of general purpose I/O and does not use the Front Side Bus or any other type of general purpose...

@Dog uhhh, the first test is using a different server than the second test, and it might be crossing between hetzner datacenters

that probably makes a much larger difference than browser choice

also I doubt most people see more than 150 - 200 Mbps in either direction on a regular basis, so having to optimize the browser to handle much more throughput than that is not a very pressing issue IMO

even if you have a gigabit connection, there's not much you could want to download that would take way less time with a gigabit compared to 200 or 300 Mbps

if I could upload video to youtube at 359 Mbps I wouldn't have to wait more than 60 seconds to upload even like an hour long 1080p video

Dog

1:01 PM

@allquicatic Same server:

Also:

Chrome, vs Edge:

Also Chrome:

vs. IE11:

allquicatic

the download speeds don't seem horribly slower, but the upload does for some reason

what data is it uploading? zeroes? is it gzip compressed? or is it RNG data?

Dog

@allquicatic Yeah that's the thing

Chrome has some weird behaviour on the upload test

tereško

what about firefox (dev edition)?

Dog

@tereško Probably closer to the IE/Edge results

allquicatic

if it's RNG data, Chrome could just be using a more expensive, (possibly more secure?) PRNG

Dog

1:05 PM

The weird upload is not something I remember seeing before, but pretty much every time over the past five years or so, Chrome has always performed slower than both IE and Firefox on these tests

allquicatic

because they're probably not buffering much of it at a time

Dog

@allquicatic The downloads are still significantly slower though - 50-100Mbps slower

allquicatic

if you wanted to buffer 5 seconds of that data you'd have to reserve 437 MB of system memory in the Chrome process's JavaScript interpreter

Dog

Ironically, Chrome is using far more memory than IE doing the test

tereško

of course it is

Dog

1:07 PM

(Also ironically, while IE is giving a much faster result it's filled with ads saying "Get a faster browser: Download Chrome!")

allquicatic

are you running this in a VM?

Dog

Maybe it is time I switched back to Firefox

@allquicatic Right now, yes, but the same behaviour was on my work PC last year (Xeon desktop/workstation)

Two different VMs in two different datacentres right now

Erf, ST beta is still broken

That was Firefox, but using the HTML5 test because the Flash test won't run

allquicatic

@Dog could be Flash on Chrome via PPAPI just sucks; did you try uploading a file to a nearby server, or even via loopback? might be more informative

setup nginx on localhost and create a random 10 GB file

Dog

@allquicatic It is a lot faster using HTML5:

But still way behind Edge:

Managed to beat Firefox's upload though

@allquicatic Eh I did a loopback with a fake Speedtest server a few years ago, Chrome could only manage 1.2Gbps, while Edge managed 1.5Gbps.

That was on my old Intel 5-series PC though :-P

It still had a northbridge

djsmiley2k

superuser.com/posts/7269/revisions

so that post... all the spams

WHATEVER Dave

1:21 PM

does speedtest use concurrent sessions to generate these counters?

perhaps chrome has a session limitation of some sort

Dog

@WHATEVERDave Yes

At least 4, up to 10 IIRC

Somehow the HTML5 site just gives a redirect loop in IE11

WHATEVER Dave

looks like firefox has a 6limit and chrome has a 6 limit for concurrent sessions

probably related

Boris_yo

WHATEVER Dave

blog.olamisan.com/max-parallel-http-connections-in-a-browser

Boris_yo

I can't edit monthly range.

I used to do that using button with 3 dots on right hand top corner.

My monthly billing range is between 2 of month until 1st of next month.

Dog

1:26 PM

@WHATEVERDave Hmm, I wonder

Chrome manages to max the upload on my other box, but Firefox maxes both the up and down:

WHATEVER Dave

ie: has 13

interesting

Dog

WHATEVER Dave

browserscope.org/?category=network&v=top <-- more accurate site

Dog

Wonder if I can find a single-threaded test site still

WHATEVER Dave

engineerhammad.blogspot.com/2015/04/Download-Test-Files.html

^-- just download one of those.. or google 1mb.txt 5mb.txt etc

;)

Dog

1:29 PM

Lol fails

Bob

@Dog @allquicatic It probably uses the Fetch API, which is still kinda new-ish and not widely used.

So varying perf optimisations.

Dog

Dark green = single threaded, light green = multithreaded

First was chrome, then Firefox

Dat stability in Chrome... not

WHATEVER Dave

Not real sure if those txt files contain binary compressed data or not

:/

Dog

@Bob HTML5 or Flash?

allquicatic

it's almost 100 Mbps faster in the upstream though

Dog

1:31 PM

I know the Flash test uses websockets

Bob

@Dog HTML5

Dog

@allquicatic Yay, 1 in 20!

Bob

@Dog Oh, websockets. Forgot about those. Yea, that makes more sense than Fetch

Dog

@WHATEVERDave Lol fail

WHATEVER Dave

anyhow, gotta meeting!.. =)

Dog

1:33 PM

I posted the link as a URL into Firefox, it tried rendering the 1GB.db as a text file

On the plus side: It used 100% of 3 CPU cores (yay multithreading!). Downside: I still had to kill it via task manager

WHATEVER Dave

lmfao

wow

Bob

@allquicatic Nope, Chrome, Firefox and Safari all use xorshift128+ now.

Also, the previous Math.random algorithm in Chrome was terrible.

Far from anything resembling secure.

Dog

@WHATEVERDave Chrome took about 20% longer (35 seconds vs 29 seconds) than Firefox

Noticeably different throughput too, left is FF, right is Chrome

Bob

@allquicatic Firefox was some old LFSR variant... Chrome used MWC1616. Both xorshift128+ now

Burgi

oops

forgot to change the password on the db connection

Dog

1:40 PM

IE bugged out and got stuck at 35Mbps

@Burgi Famous last words

See, I told you Bob knew everything

EVERYTHING

diply.net/celebs-same-age?publisher=curious

Dog

1:57 PM

> You will be sharing with me and my wee boy who is 8, 1 cat and a snail.

!!WAT

Boris_yo

How can you modify file before you create it?

Burgi

how weird...

the script executes in 1 second on the staging server, on live it times out after 300 seconds...

Dog

2:33 PM

@Boris_yo Modify a file, then create a copy elsewhere

DavidPostill

8

Q: Can a file appear to be modified before it was created?

Is it possible to have a file which has a modified date that is before its created date? What actions would cause this event to occur?

@Boris_yo

Dog

@Burgi You really enjoy having a messed up staging/production system that never behaves alike don't you

DavidPostill

Burgi

@Dog these are shitty legacy systems

it doesn't help that the website is over 6GB

djsmiley2k

@Dog one of my old jobs had 3 'pre-live' stages

all were different :/

@Burgi what can't it connect to?

Burgi

2:47 PM

its connecting but taking FOREVER to run the query

djsmiley2k

yey bad queries

your test db has 10 customers

you real one has 22 million?

Burgi

the query hasn't changed

djsmiley2k

no

the database has

Burgi

the databases are almost identical

i did a snapshot 2 weeks ago

djsmiley2k

almost as in nothing is hitting hte test one, and the live one is hammer 99.9% of the time?

I don't knwo your setup, but i'd be looknig at the differences for issues

Burgi

2:51 PM

1 sec

Dog

That's what ~~she~~ your database said!

caub

3:04 PM

There is some thin plastic thing on the cpu and gpu (in orange) i.imgur.com/nsjndAc.jpg should I remove it for cleaning more or it's fine?

there's still a bit of paste under it

Burgi

@djsmiley2k there were 300 entries in both databases

the server specs are roughly the same

however mysql/php is newer on the testing server

but not by much

djsmiley2k

weird

@caub I think that's to stop the heatsink tipping

DavidPostill

@djsmiley2k "tipping"?

djsmiley2k

when trying to fasten it down

so you don't chip the edge of the errr... chip?

like that.

caub

3:21 PM

wow

how did this happen?

I think it's dead now

djsmiley2k

no idea, found it on google

but someone likely pushed the heatsink down too much, and not straight

so it tilted and cracked the corners of the sheild

Burgi

this is brilliant:

6

Q: The Union Ain't Jack

Each line is a clue to a different word, take each of the n'th letters to get a final word that relates all the clues. Have a picture!

word visual rebus

Bob

...

Java installer: "do you want this search toolbar *cough* DEFINITELY NOT MALWARE NOPE NOSIREE *cough*??????"
me: "no, continue"
Java installer: [popup] "are you sure? we really recommend it! try selecting it! here, lemme toss you back on the same page so you can select it and click next again"

...

seriously

Oracle. Go die in a fucking fire.

Journeyman Geek

0_0

@caub It's kapton tape

I don't know if its standard but I don't think so. Feels a lot like someone used it to keep thermal compound off the non headspreader bits

DavidPostill

3:37 PM

@Bob Huh? Where exactly did you download the installer from? I've never seen this problem ...

Bob

@DavidPostill It wasn't even an installer. It was the updater.

^ that ridiculous piece of shit

DavidPostill

@Bob Ah. I never use the updater ...

Bob

It's not enough that it has to piss me off with this crapware; it has the fucking gall to ask me twice in a row.

@DavidPostill I rarely update. Decided why the hell not, this time.

Never again.

Journeyman Geek

0_0

Bob

In fact, I should just uninstall it. Had to install it to get ADB working a bit ago.

Journeyman Geek

3:39 PM

what the hell is theoma?

Bob

@JourneymanGeek A flaming piece of shit that should be sued/hacked/whatevered out of existence.

Journeyman Geek

ah. Ask

Yanno, I remember when they weren't a flaming pile of crap?

Bob

They even bloody officially documented the damn thing, including the additional prompt and need to click next a second time.

This is completely ridiculous.

Why do people still freakin' use Java.

DavidPostill

"Option 1: Disable Sponsors in the Java Control Panel (Recommended method)"

java.com/en/download/faq/disable_offers.xml

@bob

Bob

@DavidPostill Yea, I probably did that at some point on my old system.

But this has pissed me off to the point where I'd rather just purge the entire damn thing.

DavidPostill

3:44 PM

But yeah, it sucks ...

Journeyman Geek

heh

"enterprise software"

We had to downgrade office to 32 bit for the new important bit of software we're installing. Which only runs in windows 64

DavidPostill

lol

Journeyman Geek

I don't think its the app itself

DavidPostill

@JourneymanGeek Offensive comments on superuser.com/questions/1139792/no-mountable-file-systems - for some reason my flagging doesn't delete the last two ...

caub

@JourneymanGeek it was there initially, first time this laptop is opened

I don't have Java installed on my laptop, since at 2 years

Bob

4:13 PM

You forgot bandsaws, which are primarily used for cutting off your fingers. — Emily Oct 20 '14 at 18:02

@caub I wish I didn't.

allquicatic

4:27 PM

hi

Bob

@allquicatic hi