I've never seen 25 megabits from 54G, more like 12. Also, many installations I've tested show a 1-2% packetloss and upwards of 8ms of latency.. I don't see how that's acceptable for a high end residential internet connection..
You know, you could hire a professional to do it for you. It would add resale value to your house, and any little time it takes to set that up would be saved later on in terms of failed downloads, bizarre base station connectivity problems, or getting that damned child porn downloader off of your AP which has a horribly broken WPA2 implementation...
Actually, unless your ISP is hitting you with false data of some sort, your 6mbps connection is definitely faster than 802.11b.
The effective throughput of wifi is only a fraction of it's media speed - on the order of ten to twenty percent. It's also adding measurable latency to the connection. And upwards of a few percent of packetloss, in most installations.
A 10 mbps connection should be able to transfer at 1,250 short kilobytes per second minus overhead. Not 125 or 250.
Granted, you're better off just wiring quality cat5e. You can run 10/100/1000 megabits over it, with a strong guarantee of future proofing against upgrades to your internet connection (ADSL2+, that is, ITU G.992.5, can handle 24 megabits/sec in the highest profile for instance), as well as the ability to hook in devices like NAS/fileservers and such. Also you don't have to worry about the fact that your legacy B gear is running WEP, which someone else is using to access your 6mbps connection to download something the FBI is very interested in right now.
Ah if it's a continual focus/ai-servo it will definitely need a delta-time to figure out the rate that the subject is moving towards or away from the film plane.
I'm not sure that a cameraphone would have ai-servo mode though....
Other processes in the focusing system would require delays or such as well, and they might not have a convenient sleep() type call available to them for a variety of reasons (the execution context may prevent calling it, it may not be available, performance of said call might be poor, etcetc)..
In DOS or Windows environments, you can hold down alt, type in a number to directly input an "extended ASCII" or Unicode character, respectively. alt-254 under a DOS environment will give you that funny IBM-extended-ASCII square, for instance.
The IT people at my highschool so many years ago couldn't figure out how I was doing CTRL-C when they disabled the control keys on the keyboard (they put tape between the contacts).. I was just doing ALT-3....
You can buy an Intel Atom based board for about a hundred dollars with cpu, put it in a micro ATX case, throw in some discount memory, a low end, junky hard drive, and viola! Your very own mac-mini-alike thing.
Of course it will become landfill just as quickly as a mac mini as it also has no expandability, but that's okay, we wanted a mac-mini clone.
Your typical MMO is only synchronizing mobile object state between the server and client, plus handling text-chat and such. It's a constant stream, but it's low bandwidth.
I doubt Comcast does anything to prioritize that sort of traffic as it would be very hard to determine what's game traffic and what's not, as there's no standard. Some use UDP, some TCP, some both, on a variety of ports in a variety of directions with no standard headers.... The only thing they could look at is the prioritization flags in the IP header, but I don't think anybody sets or uses those...
I dunno, that sounds like just more overhead to me. We seriously don't need any additional overhead. Also, having "My Documents" or "~" under a UAC-like protection would be.. annoying.
Would be fun configuring a web server too.
I've had a lot of experience with altered user/permission execution environments, and I can tell you, it's never pretty.
Also, if you see my post below, my software validation/installation process has netted a zero infection rate for more than a decade of running Windows, Amiga, and Linux software. Over TWO decades for Amiga software, which uses an MS-DOS type security model.
Just stop running the damn viruses! It would have been 0/10 if they hadn't run them!
Seriously - I've been using 95/98/NT4/NT5/NT5.1 machines for over a decade, and I've never had a single virus. My anti-virus solutions are always installed with all of the (system breaking) protection disabled, and I just run a scan occasionally. Ditto for Amiga software prior to that. Ditto for the Linux servers I run.
Simple caution when installing things and prudent use of firewalls keeps away five nines of problems. Don't torrent l33t 0-1 day w4r3Z. Don't run cracks*. Don't use sketchy peer-to-peer software. When downloading free/oss/shareware, download only trusted, well known software, and download it directly from the source. Run md5 or sha sums, just in case. Don't let any children use your machine, or friends, or other sorts of retards, err, infection vectors. Don't use HTML-enabled email clients (I'm looking at you, Outlook), or if you do, use webmail products with a safe(ish) browser.
* If you absolutely have to, use only serial-generating ones, and run those from a secured emulated environment or system that gets re-ghosted after each serial run. It's easier just to buy the software or switch to a F/OSS solution though.
A sufficiently directional antenna doesn't exist. Even if it did, it wouldn't be able to track with the vibrations from the power fan in the system moving the wire ever so faintly as it would be TOO directional. Plus, it's in a bundle of about twenty lines. Good luck! (especially good luck as it's transmitting both ways on all eight wires simultaneously using heavy duty echo cancellation at a frequency that's a harmonic of 802.11, that'll be fun to receive on the other end! Wires, by the way, which are twisted around each other to cancel out the signal..)
The ADSL2+ signal goes to a media converter in my basement. The interceptor has to be INSIDE MY SECURITY PERIMETER, at which point they're already arrested. It's fiber optics past that point. Good luck finding a listen-only ADSL2 device. And it's SSH underneath all of that physical protection for anything remotely important. Good luck with that too!
I, am, on the other hand, surfing around the various open, WEP, or WPA based access points here (using my laptop, in default configuration, without any additional hardware, without physically visiting restricted areas). Apparently wireless isn't 'obviously less secure' enough. I'd like to see someone catch me too, if I'm just listening to packets, instead of performing B&E on telco or personal property and planting various kinds of wiretapping devices...
Beige boxing would be rather hard to pull off against ADSL2+, or my gigabit LAN. Especially the LAN since you'd have to gain physical access to my facilities first.. And every protection that can be put over wireless can be put on wired as well. Just because it's LAN/ADSL2+ doesn't mean I'm telnetting to remote servers.
In fact, if I'm going up against WiFi, I'd have to say that I'm more likely to be using something.. overheady like SSH, than some WiFi user who's wondering why their ISP "capped" them to 200k/sec, who switches to something less secure in a desperate attempt to gain back responsiveness every time the old lady next door heats something up in her microwave..
Oh and regarding the wire being a problem - I don't have much of an issue with my 50 foot ethernet cable that plugs into my notebook. It gives me full LAN speed anywhere in my unit. Also, if you're dealing with a fixed installation, a free space optics system can give you LAN-class (or superior) performance over kilometers, while acting like a wire instead of a broadcast. Full duplex, anybody?
Wired and optical technologies will ever be superior to wireless, by the simple fact that they're essentially 1D lines running through 3D space, whereas a typical wireless signal is a 3D signal in 3D space - a single frequency gives a fixed bandwidth to a single user in a given ~volume~.
Advanced tricks allow increased sharing, but the fundamental limitations remain.
Consider the volume of a typical wifi base station.. now imagine filling that volume with OC192 cabling. As they say on the "intartoobs", "pwned".
Supreme Commander was heavy threaded, yet offers hideously bad performance (even when you remove XP and Vista's terrible processor scheduling from the picture) even on core-heavy boxen.
No amount of parallelism can get around the fundamental need to serialize at some point - you have to know the result of C=A+B before you can compute C+D.
Nevermind things like oh, there's only one memory bus, additional cores is additional bus contention, additional cores is additional cache thrashing, additional cores is longer lists for CPU scheduling algorithms to traverse...
Basically, this multi-core nonsense is the CPU vendors admitting that they're stumped for innovation and need you to keep on buying so that they can have their salaries paid...
That's actually a wise upgrade path if you're not a "hArDc0Re Gam3r Do0d". However, you might want to consider buying the obsolete performance edition cards - a 6800 GTX/ultra/whatever can probably be had fairly cheaply (if not entirely free) now, and will have a lot more fill rate than say a 7300.
6800 ultra (high end, in it's time) 6.4 gigapixels/sec 35.2 gb/sec mem @256 bits 600 million vertex/sec
7300gs (entry level card from the next generation) 2.2 gigatexels/sec 6.5 gb/sec mem @64-bit 413 million vertex/sec
I ran off of a 7900GTX for a long time after it was "obsolete".. and yet most of the entry-level cards released later were (what am I saying, they still ARE) significantly slower where it really matters.
You know your rod cells are sensitive through the blue range, right? In fact, they're not sensitive in the low frequency range - a blue filter would have less effect at night than during the day as you wouldn't be losing the 'red' range (as you aren't seeing that anyways if you're relying entirely on rod cells anyways).
Vision is a logarithmic type response - losing half or three quarters of available light is only "one or two stops". If you're at an "adequate vision" status, it won't suddenly drop to "OMG IM BLIND WTFBBQ", but rather "not quite as adequate as before but still mostly adequate vision" status.
Finally, the rats (or people if so treated) would only have to worry about this if the cornea, lens, or humors of the eye became blue. The iris and sclera of the eye can be as blue as they want without impacting vision. In fact, as these structures are supposed to be opaque, the blue is actually helping!
No Problem:)
It's said that Windows 7 won't have that problem, although my own testing seems to suggest GDI performance is still sub-par. Hopefully I'm just behind on my patches or something...
This one does. The author picked the GDI approach (replacing older versions which used some DirectX) because it works better under WINE. I've not seen how it performs on Windows, but it runs well under CrossOver Games on OS X (and is highly addictive).
As I said, not many games use GDI.. I fondly remember playing "jezzball" and some other GDI-specific games back in the day, and I'm sure a number of those object-picking games are based on GDI. It's just that OpenGL / DirectX / SDL games are a lot more common...
So while NT6/7 may appear faster when you drag a solid-outline window around (Booyeah! It has AmigaDOS 2.0 features (smart windows) in it! those were CUTTING EDGE in 1989!), actual window redraws are slower. Well, on equivalent hardware. Admittedly my T7500 laptop with 2 gigs of RAM and an underclocked 8600GS with XP/NT5.1 is the same speed as my i7 with 12 gigs of RAM and GTS250 running RC1 of NT7, but that's hardly equivalent hardware.
>DX10 isn't the only reason to upgrade. SMP performance and general responsiveness is massively improved in 7 due to a better scheduler.
Sorry, no.
Win XP(NT 5.1) is better at multithreaded loads than Vista (NT 6). If you check out this thread, you'll notice in the graphs that SupCom is faster on NT 5.1 without MadBoris's optimization tool than NT 6 is WITH it. The tool statically assigns threads to cores, resulting in improvements for both OSes, suggesting that the scheduling in both is poor, with NT 6's being the poorer of the two: http://forums.gaspowered.com/viewtopic.php?p=173631#173631
IBM, Sun, and others have been trying for decades to get us to switch over to a remote execution/thin client model. They didn't succeed then, and they won't now.
It's as you said, nobody wants a single point of failure.
Correction: LCD TVs almost always have input lag. They tend to do a lot more signal processing than a typical LCD monitor.
A friend of mine got an LCD TV and hooked it to a Xbox360, with audio going through the TV, and to an external sound system through a different connection. You could actually hear the delay that the processing was adding - every sharp sound seemed to have an echo, and smooth sounds sounded... strange. If you turned off the TV's sound, the effect vanished. The TV is adding delay to the audio signal so that it's in sync with the video. And this is common for all sorts of "native resolution" type displays (ie anything not CRT) - plasma, DLP, LCD projectors, etc. It's often part of the rescaling/deinterlacing/quality enhancement system.
(Note that the absence of this effect does not mean the absence of input lag, as not all TVs will bother delaying the audio to resync it with the video)
On the other hand, SOME monitors like my own Viewsonic VX922 add little or no delay to the input. Dual head comparisons with CRTs, done by myself and others, have shown that the VX922 adds less than a frame of delay to the image, with only the occasional kick to 1 frame of delay (say 1 in 10 or less). Granted, of course, not all monitors are of this quality...but I have yet to see a non-CRT TV performing like that.
All of this makes sense when you think about it from the manufacturer's point of view.. if the audio is resync'd, does the TV viewer really care if the DVD/broadcast show is delayed by a few milliseconds? Versus monitors, which are usually used interactively, where such lag might be noticed.
Finally, regarding lag to colocated servers with your ISP: That 13ms ping you're seeing is for default size packets (56 bytes of payload for Linux, 32 for windows). Consider that this.. lunacy would NOT be sending tiny packets, but rather a heavy stream of large packets. It's NOT going to be best case latency.
We won't even discuss the utter loss of quality going from a raw 1.3 gigabit/sec signal (720p60) to heavily-compressed-with-hasty-codec 0.045 gigabit T3.
Slashdot Mirror
I've never seen 25 megabits from 54G, more like 12. Also, many installations I've tested show a 1-2% packetloss and upwards of 8ms of latency.. I don't see how that's acceptable for a high end residential internet connection..
You know, you could hire a professional to do it for you. It would add resale value to your house, and any little time it takes to set that up would be saved later on in terms of failed downloads, bizarre base station connectivity problems, or getting that damned child porn downloader off of your AP which has a horribly broken WPA2 implementation...
Actually, unless your ISP is hitting you with false data of some sort, your 6mbps connection is definitely faster than 802.11b.
The effective throughput of wifi is only a fraction of it's media speed - on the order of ten to twenty percent. It's also adding measurable latency to the connection. And upwards of a few percent of packetloss, in most installations.
A 10 mbps connection should be able to transfer at 1,250 short kilobytes per second minus overhead. Not 125 or 250.
Granted, you're better off just wiring quality cat5e. You can run 10/100/1000 megabits over it, with a strong guarantee of future proofing against upgrades to your internet connection (ADSL2+, that is, ITU G.992.5, can handle 24 megabits/sec in the highest profile for instance), as well as the ability to hook in devices like NAS/fileservers and such. Also you don't have to worry about the fact that your legacy B gear is running WEP, which someone else is using to access your 6mbps connection to download something the FBI is very interested in right now.
Ah if it's a continual focus/ai-servo it will definitely need a delta-time to figure out the rate that the subject is moving towards or away from the film plane.
I'm not sure that a cameraphone would have ai-servo mode though....
Other processes in the focusing system would require delays or such as well, and they might not have a convenient sleep() type call available to them for a variety of reasons (the execution context may prevent calling it, it may not be available, performance of said call might be poor, etcetc)..
In DOS or Windows environments, you can hold down alt, type in a number to directly input an "extended ASCII" or Unicode character, respectively. alt-254 under a DOS environment will give you that funny IBM-extended-ASCII square, for instance.
The IT people at my highschool so many years ago couldn't figure out how I was doing CTRL-C when they disabled the control keys on the keyboard (they put tape between the contacts).. I was just doing ALT-3....
Or you could just add another ComboBox. People love that shit.
The combobox is using up a lot less CPU to render. And it's code is part of the operating system, reducing your app's memory and disk overhead.
How does one "queue" awkward silence?
AddTail(&Queue,&AwkwardSilence); ?
You can buy an Intel Atom based board for about a hundred dollars with cpu, put it in a micro ATX case, throw in some discount memory, a low end, junky hard drive, and viola! Your very own mac-mini-alike thing.
Of course it will become landfill just as quickly as a mac mini as it also has no expandability, but that's okay, we wanted a mac-mini clone.
Your typical MMO is only synchronizing mobile object state between the server and client, plus handling text-chat and such. It's a constant stream, but it's low bandwidth.
I doubt Comcast does anything to prioritize that sort of traffic as it would be very hard to determine what's game traffic and what's not, as there's no standard. Some use UDP, some TCP, some both, on a variety of ports in a variety of directions with no standard headers.... The only thing they could look at is the prioritization flags in the IP header, but I don't think anybody sets or uses those...
I dunno, that sounds like just more overhead to me. We seriously don't need any additional overhead. Also, having "My Documents" or "~" under a UAC-like protection would be .. annoying.
Would be fun configuring a web server too.
I've had a lot of experience with altered user/permission execution environments, and I can tell you, it's never pretty.
Also, if you see my post below, my software validation/installation process has netted a zero infection rate for more than a decade of running Windows, Amiga, and Linux software. Over TWO decades for Amiga software, which uses an MS-DOS type security model.
Just stop running the damn viruses! It would have been 0/10 if they hadn't run them!
Seriously - I've been using 95/98/NT4/NT5/NT5.1 machines for over a decade, and I've never had a single virus. My anti-virus solutions are always installed with all of the (system breaking) protection disabled, and I just run a scan occasionally. Ditto for Amiga software prior to that. Ditto for the Linux servers I run.
Simple caution when installing things and prudent use of firewalls keeps away five nines of problems. Don't torrent l33t 0-1 day w4r3Z. Don't run cracks*. Don't use sketchy peer-to-peer software. When downloading free/oss/shareware, download only trusted, well known software, and download it directly from the source. Run md5 or sha sums, just in case. Don't let any children use your machine, or friends, or other sorts of retards, err, infection vectors. Don't use HTML-enabled email clients (I'm looking at you, Outlook), or if you do, use webmail products with a safe(ish) browser.
* If you absolutely have to, use only serial-generating ones, and run those from a secured emulated environment or system that gets re-ghosted after each serial run. It's easier just to buy the software or switch to a F/OSS solution though.
A sufficiently directional antenna doesn't exist. Even if it did, it wouldn't be able to track with the vibrations from the power fan in the system moving the wire ever so faintly as it would be TOO directional. Plus, it's in a bundle of about twenty lines. Good luck! (especially good luck as it's transmitting both ways on all eight wires simultaneously using heavy duty echo cancellation at a frequency that's a harmonic of 802.11, that'll be fun to receive on the other end! Wires, by the way, which are twisted around each other to cancel out the signal..)
The ADSL2+ signal goes to a media converter in my basement. The interceptor has to be INSIDE MY SECURITY PERIMETER, at which point they're already arrested. It's fiber optics past that point. Good luck finding a listen-only ADSL2 device. And it's SSH underneath all of that physical protection for anything remotely important. Good luck with that too!
I, am, on the other hand, surfing around the various open, WEP, or WPA based access points here (using my laptop, in default configuration, without any additional hardware, without physically visiting restricted areas). Apparently wireless isn't 'obviously less secure' enough. I'd like to see someone catch me too, if I'm just listening to packets, instead of performing B&E on telco or personal property and planting various kinds of wiretapping devices...
Beige boxing would be rather hard to pull off against ADSL2+, or my gigabit LAN. Especially the LAN since you'd have to gain physical access to my facilities first.. And every protection that can be put over wireless can be put on wired as well. Just because it's LAN/ADSL2+ doesn't mean I'm telnetting to remote servers.
In fact, if I'm going up against WiFi, I'd have to say that I'm more likely to be using something.. overheady like SSH, than some WiFi user who's wondering why their ISP "capped" them to 200k/sec, who switches to something less secure in a desperate attempt to gain back responsiveness every time the old lady next door heats something up in her microwave..
Oh and regarding the wire being a problem - I don't have much of an issue with my 50 foot ethernet cable that plugs into my notebook. It gives me full LAN speed anywhere in my unit. Also, if you're dealing with a fixed installation, a free space optics system can give you LAN-class (or superior) performance over kilometers, while acting like a wire instead of a broadcast. Full duplex, anybody?
Welcome to the real world of physics.
Wired and optical technologies will ever be superior to wireless, by the simple fact that they're essentially 1D lines running through 3D space, whereas a typical wireless signal is a 3D signal in 3D space - a single frequency gives a fixed bandwidth to a single user in a given ~volume~.
Advanced tricks allow increased sharing, but the fundamental limitations remain.
Consider the volume of a typical wifi base station .. now imagine filling that volume with OC192 cabling. As they say on the "intartoobs", "pwned".
That multithreading pipe-dream still around?
People never heard of Amdahl's Law?
Supreme Commander was heavy threaded, yet offers hideously bad performance (even when you remove XP and Vista's terrible processor scheduling from the picture) even on core-heavy boxen.
No amount of parallelism can get around the fundamental need to serialize at some point - you have to know the result of C=A+B before you can compute C+D.
Nevermind things like oh, there's only one memory bus, additional cores is additional bus contention, additional cores is additional cache thrashing, additional cores is longer lists for CPU scheduling algorithms to traverse...
Basically, this multi-core nonsense is the CPU vendors admitting that they're stumped for innovation and need you to keep on buying so that they can have their salaries paid...
That's actually a wise upgrade path if you're not a "hArDc0Re Gam3r Do0d". However, you might want to consider buying the obsolete performance edition cards - a 6800 GTX/ultra/whatever can probably be had fairly cheaply (if not entirely free) now, and will have a lot more fill rate than say a 7300.
6800 ultra (high end, in it's time)
6.4 gigapixels/sec
35.2 gb/sec mem @256 bits
600 million vertex/sec
7300gs (entry level card from the next generation)
2.2 gigatexels/sec
6.5 gb/sec mem @64-bit
413 million vertex/sec
I ran off of a 7900GTX for a long time after it was "obsolete" .. and yet most of the entry-level cards released later were (what am I saying, they still ARE) significantly slower where it really matters.
You know your rod cells are sensitive through the blue range, right? In fact, they're not sensitive in the low frequency range - a blue filter would have less effect at night than during the day as you wouldn't be losing the 'red' range (as you aren't seeing that anyways if you're relying entirely on rod cells anyways).
Vision is a logarithmic type response - losing half or three quarters of available light is only "one or two stops". If you're at an "adequate vision" status, it won't suddenly drop to "OMG IM BLIND WTFBBQ", but rather "not quite as adequate as before but still mostly adequate vision" status.
Finally, the rats (or people if so treated) would only have to worry about this if the cornea, lens, or humors of the eye became blue. The iris and sclera of the eye can be as blue as they want without impacting vision. In fact, as these structures are supposed to be opaque, the blue is actually helping!
No Problem :)
It's said that Windows 7 won't have that problem, although my own testing seems to suggest GDI performance is still sub-par. Hopefully I'm just behind on my patches or something...
Granted not many games use GDI.
This one does. The author picked the GDI approach (replacing older versions which used some DirectX) because it works better under WINE. I've not seen how it performs on Windows, but it runs well under CrossOver Games on OS X (and is highly addictive).
As I said, not many games use GDI.. I fondly remember playing "jezzball" and some other GDI-specific games back in the day, and I'm sure a number of those object-picking games are based on GDI. It's just that OpenGL / DirectX / SDL games are a lot more common...
Also, Microsoft changed the way hardware acceleration works for GDI as well:
http://blogs.msdn.com/greg_schechter/archive/2006/03/10/549310.aspx
So while NT6/7 may appear faster when you drag a solid-outline window around (Booyeah! It has AmigaDOS 2.0 features (smart windows) in it! those were CUTTING EDGE in 1989!), actual window redraws are slower. Well, on equivalent hardware. Admittedly my T7500 laptop with 2 gigs of RAM and an underclocked 8600GS with XP/NT5.1 is the same speed as my i7 with 12 gigs of RAM and GTS250 running RC1 of NT7, but that's hardly equivalent hardware.
Granted not many games use GDI...
>DX10 isn't the only reason to upgrade. SMP performance and general responsiveness is massively improved in 7 due to a better scheduler.
Sorry, no.
Win XP(NT 5.1) is better at multithreaded loads than Vista (NT 6). If you check out this thread, you'll notice in the graphs that SupCom is faster on NT 5.1 without MadBoris's optimization tool than NT 6 is WITH it. The tool statically assigns threads to cores, resulting in improvements for both OSes, suggesting that the scheduling in both is poor, with NT 6's being the poorer of the two: http://forums.gaspowered.com/viewtopic.php?p=173631#173631
[citation needed]
Most people here seem to be talking about the dynamics of payments/non-payments, loans, stolen laptops, etc.
They completely missed the point that you nailed there - this adds a whole new dimension to "pwned".
IBM, Sun, and others have been trying for decades to get us to switch over to a remote execution/thin client model. They didn't succeed then, and they won't now.
It's as you said, nobody wants a single point of failure.
Correction: LCD TVs almost always have input lag. They tend to do a lot more signal processing than a typical LCD monitor.
A friend of mine got an LCD TV and hooked it to a Xbox360, with audio going through the TV, and to an external sound system through a different connection. You could actually hear the delay that the processing was adding - every sharp sound seemed to have an echo, and smooth sounds sounded... strange. If you turned off the TV's sound, the effect vanished. The TV is adding delay to the audio signal so that it's in sync with the video. And this is common for all sorts of "native resolution" type displays (ie anything not CRT) - plasma, DLP, LCD projectors, etc. It's often part of the rescaling/deinterlacing/quality enhancement system.
(Note that the absence of this effect does not mean the absence of input lag, as not all TVs will bother delaying the audio to resync it with the video)
On the other hand, SOME monitors like my own Viewsonic VX922 add little or no delay to the input. Dual head comparisons with CRTs, done by myself and others, have shown that the VX922 adds less than a frame of delay to the image, with only the occasional kick to 1 frame of delay (say 1 in 10 or less). Granted, of course, not all monitors are of this quality...but I have yet to see a non-CRT TV performing like that.
All of this makes sense when you think about it from the manufacturer's point of view.. if the audio is resync'd, does the TV viewer really care if the DVD/broadcast show is delayed by a few milliseconds? Versus monitors, which are usually used interactively, where such lag might be noticed.
Finally, regarding lag to colocated servers with your ISP: That 13ms ping you're seeing is for default size packets (56 bytes of payload for Linux, 32 for windows). Consider that this .. lunacy would NOT be sending tiny packets, but rather a heavy stream of large packets. It's NOT going to be best case latency.
We won't even discuss the utter loss of quality going from a raw 1.3 gigabit/sec signal (720p60) to heavily-compressed-with-hasty-codec 0.045 gigabit T3.