ISP Lobbying group claims competition is bad for customers

[PlayStation 2]

Competition is bad because it's beneficial to customers?

Hey, ACA or whatever your fucking name is, America's a free market. And whoever provides what they want at what price they want wins, not your stupid ass words. Basically how I sell shit in TF2. I dunno how competition is a bad thing. Sure, too much is, but so is a monopoly.

[Trik'Stari]

1 minute ago, Dan Castellaneta said:

Competition is bad because it's beneficial to customers?

Hey, ACA or whatever your fucking name is, America's a free market. And whoever provides what they want at what price they want wins, not your stupid ass words. Basically how I sell shit in TF2. I dunno how competition is a bad thing. Sure, too much is, but so is a monopoly.

We are not a free market. The banks that are still standing after the housing market crash is proof of that.

[PlayStation 2]

1 minute ago, Trik'Stari said:

We are not a free market. The banks that are still standing after the housing market crash is proof of that.

point taken

[DevilishBooster]

Well, I wasn't going to be drinking tonight, but fuck me if these idiots don't drive me to it.....

Like, seriously? These guys are pissed off that there is a company out there that is actually listening to the market and want to expand and improve the existing infrastructure? And where are these "satisfied" customers that they claim will be "hurt" but increased competition in the networking markets? I have yet to meet one person that would say, "No, I actually really love the fact that I only have once option in my area for internet service, and their service is complete shit. I would never consider switching to a better ISP, even if it meant spending the same or a little more for higher quality service."

Where's my bottle of whiskey.....

[Trik'Stari]

12 minutes ago, DevilishBooster said:

Well, I wasn't going to be drinking tonight, but fuck me if these idiots don't drive me to it.....

Like, seriously? These guys are pissed off that there is a company out there that is actually listening to the market and want to expand and improve the existing infrastructure? And where are these "satisfied" customers that they claim will be "hurt" but increased competition in the networking markets? I have yet to meet one person that would say, "No, I actually really love the fact that I only have once option in my area for internet service, and their service is complete shit. I would never consider switching to a better ISP, even if it meant spending the same or a little more for higher quality service."

Where's my bottle of whiskey.....

It's not that there is a company that wants to spend money to make money (as far as I can tell, most corporations have lost this idea completely) but rather that the FCC is trying to force them to do so.

[Energycore]

58 minutes ago, SamStrecker said:

That isn't the main reason. There is the whole physics side of it and we have gotten so small we need other materials then silicon. 

The thing I don't get about this argument is that GPUs seem to be advancing at a much higher pace, even though they're also pushing the limits of what silicon transistors can do. Maxwell brought 25% better perf/watt (more or less) over Kepler and Pascal looks to be about the same. So did the 7000 series compared to the 6000s on the red side.

[DevilishBooster]

1 minute ago, Trik'Stari said:

It's not that there is a company that wants to spend money to make money (as far as I can tell, most corporations have lost this idea completely) but rather that the FCC is trying to force them to do so.

I have Charter as my ISP and from my experience they are always working to make their service better. On an avg day, during the busy times, I sit between high 40s Mbps and low 60s Mbps. During the slow times I'll sit at a solid high 60s Mbps. The last place I had service this good was when I had Fiber Optic from Sprint(?) out in SD.

Unless I completely missed some details somewhere, I don't see it as unreasonable for the FCC to put an expansion stipulation on a merger approval, especially for one as large as this; and from my expereinces/knowledge, Charter would have expanded/improved TWC's network area anyway, so they just kinda shrugged and said, "Uh, ok...."

[Trik'Stari]

4 minutes ago, DevilishBooster said:

I have Charter as my ISP and from my experience they are always working to make their service better. On an avg day, during the busy times, I sit between high 40s Mbps and low 60s Mbps. During the slow times I'll sit at a solid high 60s Mbps. The last place I had service this good was when I had Fiber Optic from Sprint(?) out in SD.

Unless I completely missed some details somewhere, I don't see it as unreasonable for the FCC to put an expansion stipulation on a merger approval, especially for one as large as this; and from my expereinces/knowledge, Charter would have expanded/improved TWC's network area anyway, so they just kinda shrugged and said, "Uh, ok...."

I can't speak for Charter, never had them.

 

Just Century Link. Had Verizon back a few years ago when I lived on my own. They were actually good (then). They allowed us a free upgrade from 5 to 10, which made it so that me and my roommate could game online at the same time without lag (this was in the heyday of the 360).

 

Now, I can't even solo game online with stuff like BF4 without RIDICULOUS lag, with the same connection.

[DevilishBooster]

11 minutes ago, Trik'Stari said:

I can't speak for Charter, never had them.

 

Just Century Link. Had Verizon back a few years ago when I lived on my own. They were actually good (then). They allowed us a free upgrade from 5 to 10, which made it so that me and my roommate could game online at the same time without lag (this was in the heyday of the 360).

 

Now, I can't even solo game online with stuff like BF4 without RIDICULOUS lag, with the same connection.

Yeah, I had Century Link once they were frickin' atrocious. You have my condolences. 

[thekeemo]

14 minutes ago, Trik'Stari said:

I can't speak for Charter, never had them.

 

Just Century Link. Had Verizon back a few years ago when I lived on my own. They were actually good (then). They allowed us a free upgrade from 5 to 10, which made it so that me and my roommate could game online at the same time without lag (this was in the heyday of the 360).

 

Now, I can't even solo game online with stuff like BF4 without RIDICULOUS lag, with the same connection.

That has to do with latency not bandwidth

Games are usually fairly low bandwidth

[Trik'Stari]

6 minutes ago, DevilishBooster said:

Yeah, I had Century Link once they were frickin' atrocious. You have my condolences. 

It makes it worse, that I've filed 3 FCC complaints in the last 2 years and absolutely nothing has changed.

[patrickjp93]

1 hour ago, CUDA_Cores said:

this... this doesn't even make any sense. We NEED competition in ANY market. An example of what happens if there is no competition is go look at intels CPU improvements. They've only improved a few percent a year because they have no competition from AMD.

Oh my God how many of you knuckleheads' skulls do I have to crack?!

 

Intel has plenty of competition from Sparc, PowerPC, and ARM. Every innovation Intel makes to better compete in servers gets applied to consumer parts first as a testing/validation ground. MMX, SSE, and AVX and their successors have all come into existence because of that competition, not because of AMD being competitive in the consumer space.

 

"those instructions aren't used in games so I don't care blah blah blah..." Guess what? The single-data instructions used in games today have already been optimized to their theoretical lowest clock cycle counts for the most part. These are the SISD instructions. Consumer level programmers don't make much use out of vector instructions which can multiply your performance in numerous tasks by 2, 4, and currently up to 8x with 16x on the horizon with AVX512. Intel cannot be blamed for the software market not keeping up.

 

"So why are they still selling mainstream quad-cores with K SKUs?"

Because those same 3rd rate programmers and developers suck at making multithreaded code and never learned OpenMP which makes parallel programming a walk in the park even for the brain-dead. Programs at the consumer level don't scale much beyond 2 cores, and the vast majority don't scale beyond 4. Thus, Intel fills that demand with matching hardware. For enthusiasts and professionals, Intel makes enthusiast and professional-appropriate core counts available. That said, that's a niche market, so the prices are higher because it is both low demand and low supply with more affluent buyers.

 

Anyone who blames Intel for the slowdown in PC doesn't understand the hardware, the software, nor the real fundamental problems in the industry.

[WereCat]

If they don't want to compete with other ISPs why don't they just close their business and move on instead of spreading bs?  Oh wait..their own logic doesn't apply against them selves.

[Colonel_Gerdauf]

6 minutes ago, CUDA_Cores said:

OK, so intel has a bunch of reasons why their performance gains haven't been as good that is not related to lack of competition. Then why is it back from the early 2000s all the way to 2008 we saw massive CPU performance gains from Intel and AMD? In the mainstream desktop world we don't get that anymore but in the server segment they have things like 22 core Xeon CPUs that are each faster at every single generation. Whereas the 6700K is only like 2% faster than the 4790K. Yes i understand what you are saying is the server segment is getting faster because they have good programmers writing multi-threaded applications and therefore intel keeps adding more cores for the extra demand, but why can't we see this on desktops? Why can't everyone just start using openMP like you said if it is so easy to use, then intel can start adding more cores to the desktop and giving us the performance we want? What is stopping these 3rd rate devs like you claim from using something else so intel can start adding high core-count CPUs to the desktop?

Programmers at that level have become increasingly lazy. When they see a jump in average VRAM from 2GB to 3GB for example, it tells them that they can use the extra 1GB space, to provide the same quality/performance as the would have if they spent a little extra time. We have gone from running DOOM on a Pentium P5, a marvel at the time, to needing an i5 just to run some Battlefield game, games with remarkably lackluster fidelity for the performance it requires.

[patrickjp93]

33 minutes ago, CUDA_Cores said:

OK, so intel has a bunch of reasons why their performance gains haven't been as good that is not related to lack of competition. Then why is it back from the early 2000s all the way to 2008 we saw massive CPU performance gains from Intel and AMD? In the mainstream desktop world we don't get that anymore but in the server segment they have things like 22 core Xeon CPUs that are each faster at every single generation. Whereas the 6700K is only like 2% faster than the 4790K. Yes i understand what you are saying is the server segment is getting faster because they have good programmers writing multi-threaded applications and therefore intel keeps adding more cores for the extra demand, but why can't we see this on desktops? Why can't everyone just start using openMP like you said if it is so easy to use, then intel can start adding more cores to the desktop and giving us the performance we want? What is stopping these 3rd rate devs like you claim from using something else so intel can start adding high core-count CPUs to the desktop?

1) Clockspeed boosts (the GHz war is over though, and clocks will not sharply rise again without a change in material, if ever

 

2)Because back then the SISD instructions weren't yet fully optimized. There was still room to extend the depth of the pipeline before branch miss ratios began giving negative performance gains. There was still room to refine the Out Of Order Processing engine to be able to reorder instructions in a single cycle to decrease pipeline hazards and maximize throughput. The loop detector hadn't yet been invented, circumventing the need to prefetch instructions when in small loops. Cache sizes could still expand without adding any additional cycles to a search penalty or miss penalty. Most of the general-purpose optimizations have been all but perfected in that time. Now it's down to specific cases. Programmers have to adapt to that sort of world, period. There's nothing AMD can do to push past that perfect ceiling when Intel and IBM can't. That's just the harsh reality we're dealing with.

 

Those 18-core Haswell Xeons are the biggest IC chips in the world at 662 mm sq.. The yields on them are far lower than the 150mm sq. I7 quad-cores with iGPU attached. So, what should Intel do about their pricing? Also, heterogeneous acceleration using the iGPU hasn't taken off in consumer software yet either. There's a ton of potential for increased performance on existing hardware. The problem is the programmers. There is no escaping that.

 

Consumer level programmers don't get paid enough to research better solutions, and especially in the games industry internal politics dictate that upstarts like me put up and shut up because the old guard knows better (when they most certainly do not). It takes time to teach developers a new tool. Meanwhile, existing deadlines...

 

it sucks, but that is the nature of the game. I can show you how to double the performance of existing code between a 3820 and a 4770K without having to modify much. It's just a matter of having a workload that is already properly vectorized (learning CilkPlus makes this also hilariously easy). If you want to do it via vector intrinsic functions, it's often purely a matter of changing alignment and changing 64 to 128 to 256 in the function name. And yet, if I open a game .exe file with a hex editor and start using a disassembler on it, I can translate all the hex to x86 instructions, and I still won't find anything above SSE2 even on the most recent AAA titles.

[Celmor]

Really hoping for some competition in my area, forced to one cable company, other's are offer merely DSL (country Germany).

Getting enough download bandwith but far too low upload (120/5), Latency always jittering half of the day with 20-30 ms (difference of pings) average and nice spikes (100-200 ms) so multiplayer games are a pain.

Cable company (Unitymedia) doesn't give a damn.

[BuckGup]

2 hours ago, Energycore said:

The thing I don't get about this argument is that GPUs seem to be advancing at a much higher pace, even though they're also pushing the limits of what silicon transistors can do. Maxwell brought 25% better perf/watt (more or less) over Kepler and Pascal looks to be about the same. So did the 7000 series compared to the 6000s on the red side.

GPUs are different then CPUs. Of course intel will sit around if they can. But they are still advancing forwards as much as they can. If you look at their Xeons you can see their push. 

[NoobsWeStand]

54 minutes ago, patrickjp93 said:

-snip-

 

Consumer level programmers don't get paid enough to research better solutions, and especially in the games industry internal politics dictate that upstarts like me put up and shut up because the old guard knows better (when they most certainly do not). It takes time to teach developers a new tool. Meanwhile, existing deadlines...

 

it sucks, but that is the nature of the game. I can show you how to double the performance of existing code between a 3820 and a 4770K without having to modify much. It's just a matter of having a workload that is already properly vectorized (learning CilkPlus makes this also hilariously easy). If you want to do it via vector intrinsic functions, it's often purely a matter of changing alignment and changing 64 to 128 to 256 in the function name. And yet, if I open a game .exe file with a hex editor and start using a disassembler on it, I can translate all the hex to x86 instructions, and I still won't find anything above SSE2 even on the most recent AAA titles.

 

The multiple layers of abstraction could also add to why games are not utilizing all of the resources available.

Back when everyone had to code in assembly, it was much easier to get to the hardware of the machine. Nowadays with highly abstracted code and game engines, it might not be possible for people to modify the game engine to utilize everything.

But the abstraction does come with some advantages: compatibility, readability, and accessibility. Anyone can write a c/c++ program and have it run on basically any piece of hardware they own (even some toasters). With assembly, one would have to keep in mind all of the limitations that each new device would put on them, and code around those (like some hardware specific features that would make the game run faster on some machines, but then make it incompatible with others). Code is also a lot more readable. Anyone can look at a piece of code and get the general vibe of what it might do. One would have to be a lot more dedicated to understand code that was written in assembly. There is a lot more code written by computers that actual humans; things that translate higher level languages into assembly. The code generated by computers are not meant to be read by humans. I could take Unity, and make a ball that moves to the left with a couple of lines of code. But when Unity translates that code into lower level languages, it has to add exponentially more code to get that to actually work. Why would I write tens of thousands of lines of code to get some simple ball movement, when I could just as easily do the same thing in Unity? It is just easier. So, what I think anyways, is that there will always be this gradual flow up to more abstraction.

 

There will always be a need for people who understand the lower level languages, but for time constrained things like making video games, it is just easier and more cost effective to use game engines and higher level languages

 

 

 

anyways, on topic:

Screw the little companies moaning, they should be able to compete. They are no better than the big monopolies if they complain about competition. Smaller companies can still thrive even with the pressure from the corporate overlords

 

[Colonel_Gerdauf]

1 hour ago, CUDA_Cores said:

OK, so programmers at this level are just plain lazy. What's the solution then?

 

1 hour ago, patrickjp93 said:

Consumer level programmers don't get paid enough to research better solutions, and especially in the games industry internal politics dictate that upstarts like me put up and shut up because the old guard knows better (when they most certainly do not). It takes time to teach developers a new tool. Meanwhile, existing deadlines...

 

it sucks, but that is the nature of the game. I can show you how to double the performance of existing code between a 3820 and a 4770K without having to modify much. It's just a matter of having a workload that is already properly vectorized (learning CilkPlus makes this also hilariously easy). If you want to do it via vector intrinsic functions, it's often purely a matter of changing alignment and changing 64 to 128 to 256 in the function name. And yet, if I open a game .exe file with a hex editor and start using a disassembler on it, I can translate all the hex to x86 instructions, and I still won't find anything above SSE2 even on the most recent AAA titles.

One potential solution here, albeit not a practical one, is to educate programmers-to-be with an evolving doctorate. The issue with the current education system in regards to general technology, is that once you graduate from whatever diploma you were studying for, all of the core and supplementary materials you have learned are already long past obsolete.

[matrix07012]

8 hours ago, thekeemo said:

I believe the world would be a better place if the goal wasnt profit but to help others.

\/ \/ \/

[Trik'Stari]

10 hours ago, durpflip said:

the US (assuming that's what you meant when you said "congress") barely even has any competition and prices for great internet in the US are the cheapest i've seen in my entire life. at least you don't have to deal with bigshots taking advantage of their ISP monopoly to jack prices until the upper-middle class can't reach it anymore.

$70 a month for 10 down 1 up is horrible compared to some European countries, and Asia? Forget about it.

 

The US has hardly any competition, and the companies use it as an excuse to try and do shit like data caps and throttling. Mainly because they want to lock down the internet and make it as horrible an experience as Cable or Satellite television. That's why Comcast was throttling Netflix, because Netflix was stealing their customers.

[patrickjp93]

7 hours ago, NoobsWeStand said:

 

The multiple layers of abstraction could also add to why games are not utilizing all of the resources available.

Back when everyone had to code in assembly, it was much easier to get to the hardware of the machine. Nowadays with highly abstracted code and game engines, it might not be possible for people to modify the game engine to utilize everything.

But the abstraction does come with some advantages: compatibility, readability, and accessibility. Anyone can write a c/c++ program and have it run on basically any piece of hardware they own (even some toasters). With assembly, one would have to keep in mind all of the limitations that each new device would put on them, and code around those (like some hardware specific features that would make the game run faster on some machines, but then make it incompatible with others). Code is also a lot more readable. Anyone can look at a piece of code and get the general vibe of what it might do. One would have to be a lot more dedicated to understand code that was written in assembly. There is a lot more code written by computers that actual humans; things that translate higher level languages into assembly. The code generated by computers are not meant to be read by humans. I could take Unity, and make a ball that moves to the left with a couple of lines of code. But when Unity translates that code into lower level languages, it has to add exponentially more code to get that to actually work. Why would I write tens of thousands of lines of code to get some simple ball movement, when I could just as easily do the same thing in Unity? It is just easier. So, what I think anyways, is that there will always be this gradual flow up to more abstraction.

 

There will always be a need for people who understand the lower level languages, but for time constrained things like making video games, it is just easier and more cost effective to use game engines and higher level languages

 

 

 

anyways, on topic:

Screw the little companies moaning, they should be able to compete. They are no better than the big monopolies if they complain about competition. Smaller companies can still thrive even with the pressure from the corporate overlords

 

No, the abstraction argument died a long time ago in game engines when Data-Oriented Design took over and component-based organization was utilized. Most of the fundamentals are in place anyway which is the sad part. Game engines are not highly abstracted. What you may build may seem abstracted, but the final code produced eliminates the object-oriented design and many other things automatically.

[patrickjp93]

6 hours ago, Colonel_Gerdauf said:

 

One potential solution here, albeit not a practical one, is to educate programmers-to-be with an evolving doctorate. The issue with the current education system in regards to general technology, is that once you graduate from whatever diploma you were studying for, all of the core and supplementary materials you have learned are already long past obsolete.

The only things that grow obsolete in the computer science world are hardware and specific libraries. The core of C++ 11 and 14 will never be obsolete.

 

Object-Oriented Programming is becoming obsolete for performance reasons, but organizing data and methods/functions is still useful everywhere.

 

All of programming boils down to a few simple things:

 

Data Types

Operators

If Statements

Loops

Function Calls

 

If you master those five things, you know 90% of the programming world. Sure, there's threads, there's pointers (for library writers only, as it's stupid to use pointers for much of anything in modern C++ code), there are bit-wise operations, but those are advanced things most developers have no business using. You can't make that stuff obsolete. It's just not possible.

[LukeTim]

12 hours ago, SuperShermanTanker said:

And here I am stuck with optimum 30megabits per second down and only 5 count them 5 megabits upload (the pain when uploading videos to my YouTube channel 7 hour upload when I uploaded a 12gb file of a replay of one of my live streams) and that's the only home plan they offer the only other plan they offer is business and that's freaking expensive

12GB? How long was your livestream?!

[Master Disaster]

Translation...

 

"Competition is bad for our profit as it means we have to spend more money in order to keep our customers happy ergo we either lose out (yeah right) or we bump up prices and cut investment in other areas (public services, customer support and maintenance) so we can continue to make billions a year and give as little back to anybody except our shareholders as possible"

 

Thank god that here in England at least our ISPs are somewhat decent.