Root Access

5:00 PM

but the process is what caught my eye - because I was fiddling with this stuff two months ago :P

@BenN or a butterfly

Real programmers set the universal constants at the start such that the universe evolves to contain the disk with the data they want.

@Bob You were "installing the operating system in memory, wiping the root disk and reinstalling the OS back on disk from RAM" two months ago?

But... why? It sounds like some crazy thing a hardcore geek would try for fun

djsmiley2k

@Bob I've done this with my own machines a few times.

And then there's the tale of how I once deleted mv from my system

and cp

Bob

Apr 13 at 6:30, by Bob

pivot_root worked :D

It started off as switching to a ramdisk to properly wipe the OS HDD before cancelling a server

Then it evolved to installing root-on-ZFS on a dedi server where I didn't have KVM-over-IP access

So I couldn't boot into a livecd

had to do it from a running/installed OS

allquixotic

Bob the kexec expert? :D

Bob

@allquixotic naw, kexec was a different case

that was to boot to a different kernel version in a KVM VM that itself did a kexec boot (i.e. I couldn't just point a bootloader to a new kernel)

allquixotic

5:17 PM

@Bob well, this is weird... StartCom is now offering certificates with the original business model they had, and they're apparently trusted by the major vendors, because they're signed by "Camerfirma"? O_o

Camerfirma is a Spanish company apparently authorized by the Spanish Department of Commerce

(that's their Root CA)

ssllabs.com/ssltest/analyze.html?d=startcomca.com

Bob

wtf

cross-signing shenanigans

@allquixotic would not touch with a ten-foot pole

allquixotic

it seems extremely shady that all of the sudden the Spanish Department of Commerce (or an agent/company thereof) is now cross-signing for them

and random, because StartCom is a former Israeli company owned now by China

Ben N

FWIW this is how Let's Encrypt got started

But considering the previous shenanigans of StartCom/WoSign, this is concerning

allquixotic

I think they put out an all points bulletin of sorts saying "We're going to bring money to your CA infrastructure, just let us run our business model as a reseller of your trusted CA infrastructure"

Bob

@BenN No. Cross-signing in itself is fine. Cross-signing a CA known to violate guidelines and issue fraudulent certificates, and now untrusted is not okay.

allquixotic

5:21 PM

also @Bob they're offering all SSL certs for free "for now" (with no announced end-date) including EV Class 4

free

Bob

@allquixotic Free EV? Huge alarm bells.

allquixotic

does that mean they will sell your information? lol

Ben N

Totally free, or free to generate once you pay for the verification?

Bob

EV verification is supposed to be a fairly involved process.

allquixotic

@Bob oh, they still do the verifications

you have to upload them all kinds of documents and proof of ownership of a company

but you don't pay them money for the service, "for now"

they have a price on their website but then it says it's free unless you want code-signing cert

Bob

5:22 PM

@allquixotic Yes, but one questions how thoroughly and what their (financial) incentive is.

allquixotic

@Bob Exactly, and agreed, I'm wondering if they're not using the documents people upload to create fraudulent passports, fraudulent driver's licenses, fraudulent corporate documents

2

which makes me extremely dubious to give them any info

Bob

@allquixotic tbf the majority of those documents are publicly-available

allquixotic

I mean, it's not totally out of this world for random company A to suddenly have an agreement with random company B where B lets A use their CA infrastructure to generate certificates, but the company purportedly doing the identity/company validation is StartCom

@Bob official-looking driver's licenses aren't

the numbers on them might be, but not the actual document/card

Bob

Troy Hunt documented his experience in a blogpost - it boiled down to having a (publicly-searchable) company w/ ACN (AU), plus a skype call showing a driver's licence.

allquixotic

they might be printing legit-looking ones from ones people upload to them :P

@Bob is this about StartCom?

Bob

5:25 PM

@allquixotic Yea, that's the one piece of info. Though TBH I was iffy about having to give mine to OVH.

Ben N

> until recently, Camerfirma were not re-doing the domain validations of certificates initially validated by StartCom, but Mozilla requested a change and so from next week they will be.

Bob

@allquixotic Nah, DigiCert for him

Ben N

That's from a few months back

allquixotic

@Bob StartCom makes OVH look like a peerless, 100% incorruptible, flawless company :D

Bob

lol

allquixotic

5:25 PM

OVH has only had minor issues with their president/CEO being weird

or something

ppr

still no one for this : ?

10

Q: How to get Thunderbird to nicely format Markdown?

I like Thunderbird for my email along with plain text. I disabled HTML emails and generally write most of my stuff in markdown anyway. So I'm looking for a way, to get Thunderbird to display markdown formatted plain text mails nicely when I view them. To clarify: I know Markdown Here. It's nice...

allquixotic

> Yes, somehow. Camerfirma was doing domain validations and other validations stuff for EV certs but not for OV certs and as per your suggestion they are starting doing it next week, if not earlier. They are going to do mostly everything, contacting customers, etc. We´re providing just the info for they begin contacting.

That Iñigo guy is an employee of StartCom

So StartCom is little more than a business front for Camerfirma

I guess, StartCom never gets your info, so it's all about how trustworthy Camerfirma is

I wonder what business incentive this random, never-heard-of-before Spanish company has for letting StartCom do this to them?

did they pay them money? it's not like they're making much money on the process now. They're spending company resources and time to validate peoples' info, but 99% of customers don't need code-signing so they aren't paying

> we have a contract with Camerfirma and never tried to hide it.

Ben N

Identity validation costs $60 and organization costs more

(source)

Speaking of code signing, what's the cheapest non-shady issuer of certs trusted for Windows Authenticode?

allquixotic

@BenN non-shady? I switched to AlphaSSL after the StartCom fiasco chased me away from them; not sure the cost but it's probably reasonable

Bob

@allquixotic That we have so many barely-known trusted roots is also concerning.

allquixotic

5:33 PM

@Bob well, yeah; not everyone can be a DigiCert or Thawte... and TBH those "big ones" are so overpriced that it's laughable; what, are they securing the CA infrastructure with cats that can only be trained by a single cat whisperer lady in Croatia who spends 4.5 years with each cat before it can be released to attack anyone who tries to issue improper certs?

I mean, for the price my cert better be gold plated bytes

user226528

Comodo was known as the inexpensive certificate authority.

allquixotic

@FleetCommand They're no longer the cheapest (reseller) though

there's definitely a race to the bottom in progress with the new demand for https

the cheaper ones don't come with "insurance" in case your cert gets broken, but then most people don't even need that

most people want to buy security, not security + insurance

but Thawte literally only sells security + insurance, bundled

user226528

The problem is when you buy a certificate and can't install it on your server because it is in the wrong format.\

Bob

@FleetCommand format is easy

it's the most trivial part of the process

djsmiley2k

I know a guy who worked for OVH

in Canada

Bob

5:36 PM

@allquixotic troyhunt.com/journey-to-an-extended-validation-certificate

djsmiley2k

Not perticularly nice company to work for - but then again they are basically operating on a razor edge to make profits so...

user226528

@Bob How easy? You know of a conversion tool or something?

djsmiley2k

@allquixotic oh god

allquixotic

@djsmiley2k Basically the Amazon of retail hosting, then

djsmiley2k

I want to start a SSL reseller, but the Monster Cables of resllers

Bob

5:38 PM

@FleetCommand openssl will convert from basically every format to any format

djsmiley2k

gold plated diamond enriched uranium dipped SSL certs

Bob

though the openssl syntax sucks

djsmiley2k

@allquixotic pretty much

Also, GIVE ME THE STARS!

TONIGHT WE SHALL FEAST, UPON STARS!

allquixotic

@djsmiley2k 10000x rounds of ROT26 encryption on the certificates to be 10000x harder to hack your website? for $10000 per domain?

djsmiley2k

@allquixotic $10,000 per query

These certificates are hand written, every byte

Bob

5:39 PM

@allquixotic a lotof the little-known ones are also related to at least one government

djsmiley2k

by Mongolian Pitbulls

Bob

which raises questions about surveillance

allquixotic

@Bob How so? The CA doesn't know your private key.

They'd know that you have a cert for domain X, but the company-based ones would know that too.

user226528

Aw shucks. Comodo no longer sells code-signing certificates.

Bob

@allquixotic The CA can issue certificates (in the absence of PKP) to any domain they want.

allquixotic

5:41 PM

@Bob But doing so without a customer demanding it is a severe violation of trust that led to StartCom being untrusted...

Bob

Say US gov (NSA) wants to intercept traffic. They have control over the physical network. They also have control over <government CA>. They issue a cert for <site> and, boom, MitM

allquixotic

(They did it for GitHub)

Bob

@allquixotic That's assuming it's detected.

allquixotic

@Bob Well, check for any CAs you can buy certs from that have a root CA of the US Department of Defense, Department of State, etc. and avoid like the plague? :P

djsmiley2k

@Bob 1. they walk into your office

2. gag

Bob

5:43 PM

If it's not published in certificate transparency logs (which startcom/wosign did), and it's used in fairly narrow scenarios, no one will find out.

allquixotic

the Five Eyes doesn't include Spain, IIRC.

djsmiley2k

3. break kneecaps

4. PROFIT!!!

Bob

@allquixotic Problem isn't if you buy certs from them.

allquixotic

Ah, I see what you mean

Bob

I can use LE certs, and <government CA> could still decide to sign a cert for my domain. Which my browser would still trust.

djsmiley2k

5:44 PM

switcharoo

Bob

If they used it broadly, grc, etc, will find out.

djsmiley2k

Only solution is remove goverment CA from your trusted store.

trust no one

verify everything

allquixotic

@Bob thank god for HPKP and PKPK and HJHJ and JKJK and FOURLETTERSWITHTHEFIRSTTWOREPEATED*infinity, then? :D

Bob

But if they just target one small location... most people would never realise.

@allquixotic yea, that's why those things exist

specifially because of the threat of a rogue CA

allquixotic

I suppose that would be a thing the US govt would do... or Russia... or China...

Bob

5:46 PM

@allquixotic That's why the wosign thing was so concerning. Because China's gov might've been involved.

Even private companies are, obviously, not immune to this.

djsmiley2k

or Korea...

user226528

North or south?

Bob

but a CA directly operated by a government is a bit more ehhhh

allquixotic

@Bob I'm not predisposed to trust a government any more or less than a company. Both have a lot to lose if they betray trust and get caught, and a lot of reasons why they wouldn't want to do that unless they had something so critical they were willing to take a huge risk and throw morals out the window.

user226528

True dat.

allquixotic

5:48 PM

If the Chinese government were involved, it would undermine their diplomatic trustworthiness on the world stage significantly

user226528

Surprisingly, they both care and don't care at the same times.

allquixotic

Also, consider that some governments at least have the illusion of being ruled by popular consensus of their people, whereas all corporations (that I know of) are effectively small totalitarian regimes operated by a dictator with a small oligarchy/cabinet informing his or her decision-making.

Bob

@allquixotic OTOH most governments are in a better position to actually perform a MitM

user226528

But they also have more to lose.

Bob

@FleetCommand Debatable. The NSA/Snowden revelations didn't affect them that much, did they?

allquixotic

5:52 PM

@Bob Under the hood, it affected them tremendously and actually created a cost center in the black projects that I speculate probably exceeds $10 Billion US

You can find out from public information that that's the case

user226528

@Bob In that certain case, they weren't performing MitM; they were spying on foreign nations. Even if that's not entirely true, it is the pretext behind which they hide.

user226528

@Bob Also, I can argue that U.S. dreads that happening again.

Bob

@allquixotic Mhm, but did it do much to stop them, tone down the projects? Though that'd hardly be public info.

user226528

It is the U.S. They are set in their ways. They don't change.

Bob

@FleetCommand Yea, the chance of detection in this case is considerably higher than some other attacks so it's unlikely they'd use this. I'm just saying the possibility is there, depending on how much the gov cares.

China's already doing a number by artificially throttling encrypted communications. Public image might've been the only thing stopping them, not that they had a particularly good public image in this regard.

user226528

5:57 PM

Remember President Bush? He declared a group of countries as the axis of evil. Other presidents did a similar things too, under other, more euphemistic titles. But POTUS has ever declared an axis of angel. Do you know why?

djsmiley2k

@allquixotic and then there's Brexit.

Bob

anyway, naptime. 'night

allquixotic

nini Bob

user226528

Holy crap, nobody told me an NDA could be 25 pages! I thought half a page at most.

Avery

half a page is too short, 25 is too long

user226528

6:01 PM

Oh, wait. It is forbidding me from disclosing something that I already disclosed. (Sigh!)

user226528

I always thought a Japanese NDA would be three words long: Mizaru, Kikazaru, Iwazaru.

user226528

Or, three characters, because there are Unicode characters for these three words:🙈, 🙉 and 🙊

djsmiley2k

ur not allowed to say u've got a NDA?

are you allowed to confirm you're not allowed to say it?

are you forbidden from denying you are forbidden from confirming a nda?

Ben N

6:34 PM

Heh, just got my first 4K display; that resolution makes it abundantly clear which applications are DPI-aware

For example, the Office setup window is sized reasonably, but it's mostly blank because all the UI elements are tiny

allquixotic

@BenN the only experience I have with HiDPI is "Retina" displays on Macbook Pros that, especially on Windows in Bootcamp or with Wine, sometimes have to render at the full pixel resolution of the display

even though on MacOS, native Mac programs do a form of supersampling and scaling

Ben N

Windows can do the scaling for programs, but apparently some claim to handle high DPI and then do it wrong :|

allquixotic

you think that's bad, try handling HiDPI with DirectX 9 games

Ben N

I should probably test my own programs on this screen, huh

allquixotic

I play a few DX9 games and there's no end of resolution problems with my ultrawide display and HiDPI displays

things sized wrong, aspect ratio wrong, blah blah

games targeting a more modern API are more likely to properly handle res

Ben N

6:40 PM

How wide is ultrawide? I'm not sure I've ever experienced such a display

allquixotic

@BenN acer.com/ac/en/US/content/predator-x34-series

also the "built-in overclocking" to 100 Hz G-Sync is legit; it's (1) stable, (2) smooth (extremely smooth), and (3) does in fact hit and stay at 100 Hz if your GPU can keep up

Ben N

Impressive

allquixotic

it's not the greatest at the browser blur test, but some of that could be browser rendering latency, that might not be present in DX12 games

blurbusters.com

testufo.com/#test=framerates

Ben N

"Browser stutter" detected - I guess I have too many tabs

Though I also have Excel, IDA, VS, and a few lighter apps going too

allquixotic

IDA Pro? whoa

don't you have to, like, sell them a car and provide them copies of your birth certificate and a pee sample to get access to that "superpowerful godly software"?

That REDACTED Guy

6:51 PM

@allquixotic That's an interesting site.

Ben N

Naw, I have the last version they let you have for free

Doesn't work on 64-bit stuff, but I just care about 16-bit programs anyway

djsmiley2k

7:10 PM

yaawn

bwDraco

roar

@allquixotic Don't be silly. IDA Pro is expensive, but not extraordinarily so. It runs just over $1,000 for a standard named-user license last I checked.

@allquixotic I hate to say this again, but I'm really envious of you. I'm still stuck waiting for renovations.

I don't think I'm going to take well to ultrawide. I'm rather used to having two displays, and I tend to rely on the second display for system monitoring while gaming.

Nzall

sigh

Looks like my internet problems that used to only happen every other Saturday now happen on other days as well

it's always the same that happens: I lose internet on all my networks, followed by the network coming back 2 minutes later at 1% of the normal speed

bwDraco

7:27 PM

On Intel's upcoming 12C+ Core i9 processors: tomshardware.com/news/…

Core i9-7920X will have a base clock frequency of 2.9 GHz. Not good.

Avery

10C one was impossible to cool with even water after some wattages

djsmiley2k

@bwDraco lol they are getting slower.

bwDraco

Intel's Skylake-SP server processors were never designed to operate at enthusiast-level clock speeds on all cores. The i9-7920 (12C/24T) and above use the Skylake-SP MCC die, which allows up to 18 cores. However, this compounds the thermal problems already present in the Skylake-SP LCC die used in the i9-7900X (10C/20T).

Not sure what the all-core Turbo Boost frequency will be.

The AMD Ryzen Threadripper 1920X (12C/24T) will operate at 3.6 GHz on all cores, boosting to 4.0 GHz.

AMD's 16C part (1950X) will run at 3.5 GHz. How AMD is hitting these kinds of clocks is beyond me.

djsmiley2k

long ago Intel decided to run hot and now it's biting them in the ass ;D

long ago, as in when they pilled the pressure onto AMD way back wen.

bwDraco

To add insult to injury, AMD has set extremely aggressive pricing for these processors. The 1920X is priced at $799 when the i9-7920X will sell for $1,189.

The 1950X will cost $999, while the i9-7960X (16C/32T; clocks unknown) will sell for $1,699.

It remains to be seen how well AMD's processors will scale because of the MCM design, but early benchmarks with 32C/64T EPYC server processors paint a decent picture.

That REDACTED Guy

7:39 PM

@Nzall Call your ISP. Or change your ISP. Or threaten to change ISPs.

allquixotic

@ThatREDACTEDGuy sucks when you have no good alternatives though

bwDraco

No matter how you look at it, AMD is putting some very serious heat on Intel and they're having trouble competing.

Nzall

@ThatREDACTEDGuy Already contacted my ISP via Twitter. I live in Belgium, and while both major ISPs here have coverage and decent speeds, A) Proximus is still slower because they're over the phone, not the cable, and B) rebooting the router fixes it

allquixotic

here we have:
3.5 Mbps ADSL that drops out when it rains, which at some parts of the year is maddeningly often
advertised 100 Mbps down / 1 Mbps up Coax Cable, with actual speeds usually no better than 40 Mbps down / 512 up, with broken routers that reset the entire cable connection when you try to upload anything more than an HTTP request, and drops out when it rains or a big truck drives down the street
(in 32 days!!!!!) Verizon FiOS fiber to the premises, which hopefully will be reliable :D
unlimited 4G LTE on AT&T, which is rock-solid reliable and portable, but speeds can dip as low as …

Nzall

@allquixotic I have 200 down and 20 up, which I actually get, or close to according to speedtest

bwDraco

7:47 PM

@allquixotic congrats on finally getting Fios \o/

allquixotic

I'll be getting 150 down / 150 up (yay symmetrical, awesome for youtube streaming) in mid-August

I'd rather have symmetrical than double the downstream with pathetic upstream.

bwDraco

I'm on 100/10 Charter Spectrum (formerly Time Warner Cable) here in New York.

allquixotic

@bwDraco it's been how many weeks since 2007? :) they told us we'd have it in "weeks" in 2007

bwDraco

Never really felt the need for super-fast upstream speeds.

allquixotic

@bwDraco when I take a screenshot with puu.sh, or need to upload a 7z to my server for dev, or whatever, it's awesome to have decent upstream

also system backups to the cloud, iCloud, etc

bwDraco

7:49 PM

Yeah, I know, but I don't do these things often.

allquixotic

LTE upstream is second only to FiOS; it's surprisingly decent off-peak

can hit 25 Mbps at home

bwDraco

We've always had Fios in this area. That Verizon took so long is very, very shameful of them.

T-Mobile LTE is really, really fast around here as long as you have a capable device.

allquixotic

@bwDraco NYC is worse; Verizon got federal funding to build out FiOS in NYC, then exploited contract verbiage to say they covered people by merely passing their house or passing the street they live on with fiber, even if not delivering/offering service to them

bwDraco

Their service still sucks in less-populated areas, but that's changing fast.

allquixotic

in my zip code they were supposed to roll out to everyone, but only certain advantaged neighborhoods got it until just now

That REDACTED Guy

7:51 PM

@allquixotic You're talking to someone who has only ONE option. Whatever trick my ISP decides to play on me, I can either endure it or live without internet.

bwDraco

@allquixotic I think they got sued for this?

allquixotic

@bwDraco I think they may have, but can't recall if they won or lost

and if they lost they probably will appeal until 2096

it's a fact that rolling out a new landline connection to a premises can be extremely expensive, beyond just the cost of the cabling and infrastructure

the big cost centers would come from digging new right of way (or tunneling a cable through the earth), etc. or getting access to the existing right of way which can be expensive work in and of itself

calling a company to ensure that any digging doesn't damage existing buried infrastructure

and I think Verizon thinks (whether true or false) that people don't want the service, so they aren't going to spend the money unless people want it

djsmiley2k

sigh

i keep reading more about quantum entanglement

and I can't help feeling that a whole lot of the stuff we've been told is all hoodoo

DavidPostill

8:14 PM

@djsmiley2k Hmm

Hoodoo (geology)

A hoodoo (also called a tent rock, fairy chimney or earth pyramid) is a tall, thin spire of rock that protrudes from the bottom of an arid drainage basin or badland. Hoodoos, which may range from 1.5 to 45 metres (4.9 to 147.6 ft), typically consist of relatively soft rock topped by harder, less easily eroded stone that protects each column from the elements. They generally form within sedimentary rock and volcanic rock formations. Hoodoos are found mainly in the desert in dry, hot areas. In common usage, the difference between hoodoos and pinnacles (or spires) is that hoodoos have a variab...

:D

@djsmiley2k Ggbo

o_O?

> Hoodoo, known as “Ggbo” in West Africa, is African-American folk magic.

djsmiley2k

:OOO

DavidPostill

8:16 PM

The Difference Between Hoodoo And Voodoo

allquixotic

@djsmiley2k A lot of people have trouble believing that information can propagate at faster than the speed of light, which is what QE supposedly enables. It appears to violate a lot of fundamental principles of physics.

djsmiley2k

@allquixotic I can't find anything to say that it ACTUALLY enables it

as you can't control the state, only view it

allquixotic

There's a bit of a "crisis" in the physics world that has been widening over time as quantum mechanics and the rest of the non-quantum standard model continue to diverge, and both have experimental proof to back them up.

djsmiley2k

I've tried to figure a way around that problem out, but still, I can't

so the only infomation moving FTL is the state itself

which you can't control.

allquixotic

@djsmiley2k Light takes between 3 and 22 minutes to travel between Earth and Mars in a vacuum, depending on the orbital phase of Earth and Mars relative to one another. To really prove out QE propagating information, you'd have to communicate the most simple message possible -- "Now!" -- in less time than that, while keeping two atomic clocks synchronized at both Earth and Mars, and have a computer activate something on Mars resulting from that message.

The message "Now!" is a message indicating, effectively, "Do ____ when you receive some information matching a given set of physical characteristics" -- it's super, super simple and can be represented in binary as a 1 digit.

So if you could send that message on Earth at a given time and have the reaction to that message occur in less than the lightspeed distance between Earth and Mars, you'd have proven that QE works.

And the time differential is enough there to be sure it's not rounding.

You could have a "race" to turn on two separate LEDs, with two computer systems, one based on a message through QE, and one based on a classical lightspeed microwave signal -- the system simply illuminates the LED as fast as possible once it receives the message.

on Earth the delay is so low between two points on Earth that it's quite hard to get the parameters accurate enough, and there are too many places where experimental error could make the experiment appear convincing when in fact something is borked

Michael Frank

8:32 PM

So how about all them phones that are rebooting when people try to call 911?!

allquixotic

but if FTL QE is true, that 3 minutes should collapse into sub-1-second times between Earth and Mars

Ben N

No-communication theorem

In physics, the no-communication theorem is a no-go theorem from quantum information theory which states that, during measurement of an entangled quantum state, it is not possible for one observer, by making a measurement of a subsystem of the total state, to communicate information to another observer. The theorem is important because, in quantum mechanics, quantum entanglement is an effect by which certain widely separated events can be correlated in ways that suggest the possibility of instantaneous communication. The no-communication theorem gives conditions under which such transfer of...

DavidPostill

@allquixotic Surely that is already testable. Using the next whatever probe when it reaches sufficiently far from earth. Binary 1 received blows it up or stops it blowing up at a preset time or similar.

allquixotic

@DavidPostill yes, it's surely testable, even without a manned presence on Mars, with a rover, or with a probe that has QE capable hardware on board

it's testable in a laboratory, too, but the amount of parameters you have to account for to say "hey, by the speed of light it should've taken 18 nanoseconds, but it actually took 16!" is pretty dubious due to the huge sources of experimental error that are possible

djsmiley2k

@allquixotic I don't think you understand this infomation thing

the photos have 4 states

you don't get to decide the states

you go 'photon, show me your states', it goes "UP, One, Blue, Flat"

allquixotic

8:37 PM

@djsmiley2k I don't think China does, either, since they're saying they accomplished FTL with a satellite recently

djsmiley2k

now you know, the photon on Mars is also showing "Up, One, Blue, Flat"

Tell me how you move information using this, (other than finding out the state of the proton, which you could of done anyway).

Did they claim that?

I just saw them claim they proved entanglement over the longest distance so far.

@DavidPostill you can't ask the photons state

otherwise you set the state

So, lets say 'Spaceship hitting mars causes the state to be decided by viewing the photon'

Ok, 2nd photon is on earth.

How do you know it's states been decided?

NotAdmin Dave

grafana = pretty freekin awesome

djsmiley2k

If you observe the photon, you've changed it

and so the one on the spaceship (which hasn't actually yet crashed) is now set.

This is why I have doubts over this chinese encyption.

Avery

8:57 PM

@NotAdminDave yup

NotAdmin Dave

just got it up and working

still figuring it out

:D

djsmiley2k

ahhhhhhhhhhh that thing

brings back memories

hardest thing with any kind of monitoring, is knowing what you actually need to monitor (in a business situation)

NotAdmin Dave

yup ;]

djsmiley2k

at home, go crazy,

and setting the correct thresholds for things

NotAdmin Dave

esxi monitoring

and synology

djsmiley2k

8:59 PM

:)

@NotAdminDave is this at work?

NotAdmin Dave

the first image is yeah

second one i found on the internet and passed it off to our esxi guy

djsmiley2k

;D

if you want someone to write nagios checks for you....

NotAdmin Dave

:D

Burgi

9:40 PM

@djsmiley2k £60

allquixotic

@djsmiley2k Fair enough; but if it could be demonstrated that communication of actual, actionable information were possible at speeds greater than c, it would create a crisis in physics

I'm just saying what the parameters of the experiment would have to be for any FTL experiment to be testable, and that doing it on earth or even near earth orbit is too small a distance to get that warm fuzzy that there isn't any way experimental error could be involved

tereško

hmm .. guys

would it be reasonable to put a 450W PSU in a GTX1070 + R7 1700 build?

allquixotic

@tereško 95W + 150W for the CPU + GPU, then add another 30 to 60 W for your motherboard (USB, PCH, PCIe slot power, etc.), and you're looking at about 305W. 450 is decent headroom unless you're using zillions of USB devices that aren't connected to powered hubs, or 8 HDDs or something

bwDraco

Corrections: AMD Ryzen Threadripper 1920X runs at 3.5 GHz, not 3.6 GHz base. 1950X runs at 3.4 GHz.

allquixotic

it'd probably work even if the PSU was designed cheaply and couldn't really hold 450W efficiently or for long periods, because I don't think you'd ever hit 450W in a real world workload

bwDraco

9:51 PM

Regardless, Intel is having lots of trouble with getting those Skylake-X cores to run fast enough without running into serious thermal issues.

tereško

I am looking at x50W units, because those tend to be from SeaSonic OEM (at the low wattage ranges)

allquixotic

oh, if you're buying Seasonic I don't think you could break your PSU even if you exceed its advertised wattage; they're virtually the highest-quality PSU maker on the planet

they build their PSUs like Dover builds elevators: build in a ridiculous amount of over-engineering "just in case" 30 morbidly obese people try to squeeze inside at once :D

450W from seasonic is super conservative; you'd be hard-pressed to see any negative effects with a steady 550W load, IMO.

tereško

dunno, mine is from SuperFlower (well, it ways EVGA on it, but I looked up the OEM)

but yeah, at low wattages the x00W PSUs tend to be from FSP

not a hard-rule but a trend

bwDraco

I'm getting Seasonic for my upcoming Astaroth desktop. PRIME Titanium no less.

reddit.com/r/Amd/comments/6o7hau/comment/dkfaxxp

> [...] this is an additional benefit of the small die strategy. Binning is much easier when there are only 8 cores per die. Where Intel has to have 28 cores on a single die that are all willing to clock at 3.5 GHz, AMD only has to find 8 cores that will do that on 4 dice.

...this small die thing is fucking genius. It wipes out Intel's greatest strength (process tech) and buries them under the raw probabilities of defect and yield rates. The best thing Intel can do in response is copy AMD and abandon every technical innovation they just baked into Skylake-SP. It's 2003 all over again. I...…

tereško

this year Intel seems to do a pretty good impression of a blackout-drunk guy on a frozen lake

bwDraco

10:04 PM

You're beaten, Intel.

►

It's actually 18 cores on the die, not 20, but AdoredTV makes an excellent point about the whole issue of manufacturing cost. One reason Intel's biggest chips are so expensive is that they are insanely difficult to manufacture. AMD totally sidesteps this issue by using multiple smaller dies and putting them together using the Infinity Fabric.

tereško

I guess we will have to wait till actual benchmarks

... but based on Mobo spec's this will be interesting

bwDraco

The AMD Zeppelin die is just under 200 mm^2 in size. AnandTech estimates a size of 473 mm^2 for Skylake-SP HCC.

That's bound to make a huge difference in cost.

(the MCC designation is now HCC; more than that, and it's XCC, extreme core count)

allquixotic

10:44 PM

@bwDraco I prefer the term YAGNICC (You Aren't Gonna Need It Core Count) for anything more than 4 cores for consumers

bwDraco

Yeah.

allquixotic

well, 6 (high single thread perf) cores for AotS and future expansive strategy games, I guess

bwDraco

Jun 4 at 7:14, by bwDraco

I could use 6C or 8C but that's a "once-in-a-while, can live without it" thing.

Avery

stackstatus.net/post/163188390264/…

Journeyman Geek

@Bob yup superuser.com/questions/523928/…

@Bob there's a passel of other issues I need to sort out

(I can't get the resolution down to what this supports over VGA, it dosen't accurately report its resolution, it renders horribad on windows, this setup seems to crash explorer.,...)

Transcript for