-

[mjohansson]

Id travel throughout time and just have a look of things, see what the dinosaurs looked like. Then Id teach people whatever I could, like how to create fire, to eat broccoli if you have a cold, to disenfect wounds with alcohol, hmm that might be the extent of my advanced 21th century knowledge... Hey mayby I could learn a thing or two from people in the past instead!

 

[T.T.]

I would go to the future and buy a really good computer, then go back in time and give it to past people, then repeat


before you know we have computer with 5000 core's running at 1 Thz

 

[Jourdy288]

I'd destroy it and the plans for it. Messing with it would be a terrible idea.

Just my thoughts.


Before doing something responsible, however, I'd probably tell Hannibal not to waste time with elephants and advise Alexander to press on.


Or, I could just travel forward five years, take a visit to the patent office

I would go to the future and buy a really good computer, then go back in time and give it to past people, then repeat


before you know we have computer with 100 core's running at 1 Yhz

Fix't.

 

[Blue Protoman]

Yeah, but then you'd have to know how to write your programs to use all of them.

 

[T.T.]

Yeah, but then you'd have to know how to write your programs to use all of them.

good point

 

[Binky]

Why not make copies of future software and bring them back on a thumbnail-drive?

 

[fusion_power]

I would travel back in time and go to epic Concerts like QEEN live at Wembley 1986 or a Elvis Concert or Beatles...

 

[diligentcircle]

Yeah, but then you'd have to know how to write your programs to use all of them.

Why would you write super-inefficient software just so you can use up everything on a future PC?


The real problem would be architecture: you would need to make sure to get one with a processor compatible with x86 or at least AMD64 or ARM, and depending on how far into the future you travel, there's no guarantee that these architectures would be supported any more.

 

[Jourdy288]

^Good point.

 

[Blue Protoman]

Yeah, but then you'd have to know how to write your programs to use all of them.

Why would you write super-inefficient software just so you can use up everything on a future PC?


The real problem would be architecture: you would need to make sure to get one with a processor compatible with x86 or at least AMD64 or ARM, and depending on how far into the future you travel, there's no guarantee that these architectures would be supported any more.

Well what if you wanted Photoshop to run as quickly as MS Paint XP? Surely that's possible with 500 cores, right?

 

[Binky]

Yeah, but then you'd have to know how to write your programs to use all of them.

Why would you write super-inefficient software just so you can use up everything on a future PC?


The real problem would be architecture: you would need to make sure to get one with a processor compatible with x86 or at least AMD64 or ARM, and depending on how far into the future you travel, there's no guarantee that these architectures would be supported any more.

Well what if you wanted Photoshop to run as quickly as MS Paint XP? Surely that's possible with 500 cores, right?

Will Photoshop run in 500 different threads?

 

[diligentcircle]

Yeah, but then you'd have to know how to write your programs to use all of them.

Why would you write super-inefficient software just so you can use up everything on a future PC?


The real problem would be architecture: you would need to make sure to get one with a processor compatible with x86 or at least AMD64 or ARM, and depending on how far into the future you travel, there's no guarantee that these architectures would be supported any more.

Well what if you wanted Photoshop to run as quickly as MS Paint XP? Surely that's possible with 500 cores, right?

If there is any real benefit to Photoshop using multiple cores, it's already set to do so or it will be.


The truth is multiprocessing is not the easiest thing in the world. You need to figure out what is actually able to be separated safely from the rest of the logic for some time, and whether or not it actually presents a benefit in speed. It's also not the most difficult thing in the world, but if there isn't a need or major benefit to using multiple processors, there's no reason for a program to do so.

 

[Blue Protoman]

Well, gimme another program that runs notoriously slowly. I think it's been established that cores themselves can't get much faster, right?

 

[Binky]

I know! Raytrace in realtime with all the bells and whistles turned on. That could take 3D games to the next level of unnecessary realisticness (is that a word?)

 

[diligentcircle]

Well, gimme another program that runs notoriously slowly. I think it's been established that cores themselves can't get much faster, right?

Yes, that's just about true AFAIK. But like I said, if there's a need for the program to run on multiple cores, it will. If it runs notoriously slowly without the aid of multiprocessing, either it will support multiprocessing soon enough or it will die a slow death, and some alternative that does support multiprocessing will pop up (or rise in popularity). On the other hand, if a good single core can run it at a good speed, then it shouldn't matter given your super-fast 500 core processor of the future, because the trend of choosing multiple cores over faster cores is already here (so I doubt single cores would get even slower in the future).

unnecessary realisticness (is that a word?)

No. The word is "realism".

 

[WizardStan]

Will Photoshop run in 500 different threads?

It will if you do it wrong.