Originally posted by willmore
View Post
Announcement
Collapse
No announcement yet.
Intel Makes Cryptography Faster On Linux
Collapse
X
-
Originally posted by willmore View PostI never said it was insufficient for my needs. I said that it could have performed better.
Originally posted by willmore View PostI'm sorry that you can't understand why a consumer would be unhappy with Intel and their business practices.
Comment
-
Originally posted by Gusar View PostYou showed the middle finger to Intel. That says more than just "they could have performed better", it says you wanted a feature but it's not there on your specific processor. It doesn't make sense to show the finger for the lack of a feature you don't need anyway.
Intel uses the exact same silicon for both SKUs, one has some of the silicon just disabled/burned off. It's a concious decision to sell basically exact same hardware at different price points.
Comment
-
Originally posted by tomato View PostIntel uses the exact same silicon for both SKUs, one has some of the silicon just disabled/burned off. It's a concious decision to sell basically exact same hardware at different price points.
Comment
-
@willmore
Gusar is right, you have bought that processor knowing its limits, why are you angry now? It would make sense if you bought a processor with that feature and it didn't work, but this isn't the case..
p.s. if you notice almost every producer sells crippled versions of its products to differentiate price ranges.
Comment
-
Originally posted by willmore View PostBecause they have a business model based on non-reality based feature differientation? Does AMD do that?
Previously Black series vs normal phenoms.
There are two methods to produce segmentation:
- produce one full featured design and cut the parts out. Top-down.
pros:
1) can mix damaged chips
2) easier to support by means of drivers etc, due to same "housing"
3) cheaper to make, because they are made en masse and cut down as needed
4) many other needed parts match all crippled models
5) as effect, longer support cycle because cheap to maintain
cons:
1) no one likes crippling
2) more raw material waste
3) more energy loss
- produce many designs, designed for each specific case. Bottom-up.
pros:
1) can pack more chips on single waffle, less raw material waste
2) more energy efficient
cons:
1) they actually have nothing to de-cripple. They are crippled from design
2) harder to maintain, smaller support window. For example, compare G92 based nvidia 8800-gtx295 vs HD2xxx-4xxx.
3) harder to test, more bugs
4) development costs more, more wastes due to incorrect overproduction
* do not perform any segmentation, only filter chips based on their readings and clocks.
Thats what I prefer.
But I don't think this should interest you.
You just pick good price and all features you need and you are done. Let manufacturers decide how they implement it - crippled after or crippled on paper, in the end it plays little role.
That was an excellent troll you postedLast edited by crazycheese; 15 December 2012, 09:42 AM.
Comment
-
Originally posted by frantaylor View PostAnd AGAIN since the instruction set is just an abstraction and has NOTHING to do with the actual hardware you might as well shoot for performance. Who cares about that dreadful CISC code? Nobody is ever going to look at it. It shoots out of the compiler and into the instruction unit and nobody needs to actually look at it or appreciate its intrinsic beauty.
You know, one of the funding principles of RISC was that compilers can't use too complex instructions.
Comment
Comment