G
Guest
Guest
First off, this is a serious post, no extremists please. This is my first post here, and I’d LOVE it if it was worth my time to set up the account.
I have a question, I hear a lot of Intel bashing, and that’s ok, they are the “Big evil company” that every one loves to hate. That’s fine, we all need one or two. But, how many people have forgotten that the P4 is a vector based processor? It is the first of it kind (to the best of my knowledge) available to the general public.
What this means exactly, I don’t know (something about 90% of your time being in a loop, and reducing the overhead for the loop). What I do know is that there are ZERO compilers available for Windows (No, not even MS’s new 7.0 compiler), and only ONE that was just release for Linux, that will actually use this.
I wish I still had the URL, but in several CAD programs (which I would assume use similar operations as video games, please correct me if I’m wrong), Intel had mad a plug-in that re-compiles the program with vector support.
In these bench marks, with this compile, it made the AMD MP perform almost 60% faster in the top 6 operation (operations used the most). This was great! But the P4 results made my jaw hit the floor (Yes, I’m an AMD user, so don’t peg me as some Intel lover.). ALMOST 3 TIMES FASTER!!! It put the P4 (in the same top 6 operation) 1.4GHz ABOVE an AMD MP 1.2 (by a large margin, from being beaten into the ground without).
I was amazed, now, I know you would never see that kind of performance boost in real life, but still, when put in a native environment, the P4 SCREAMED.
I don’t expect to see a recompile of Direct X or Windows with this compiler (sense MS is now pushing there new compiler), but it’s already available for Linux. My question is, what happens in a year or two, when this compiler is available to game companies?
Remember, the P4 is in, what I would assume, a backward compatibility mode, and NOT it’s native processing mode with current compilers.
Also, I’ve heard that AMD is having yield problems past 1.4GHz. Is this true (I’d like an answer from someone that KNOWS, and not some know-it-all jerk please)? If this is true, what happens when the P4 goes to a smaller die and can hit even HIGHER clock speeds? Also, I’ve heard that more of the ORIGIAL design for the P4 (It was scaled back remember?) will be added back in. If this is true, does that not help even more so?
In the end, I half to ask, in two, three years, when we look back at the two chip, which one will end up being faster?
I’ve been out of the loop for a while, so go easy if my info is wrong; that’s why I’m here, to get it “updated”.
Peace!
I have a question, I hear a lot of Intel bashing, and that’s ok, they are the “Big evil company” that every one loves to hate. That’s fine, we all need one or two. But, how many people have forgotten that the P4 is a vector based processor? It is the first of it kind (to the best of my knowledge) available to the general public.
What this means exactly, I don’t know (something about 90% of your time being in a loop, and reducing the overhead for the loop). What I do know is that there are ZERO compilers available for Windows (No, not even MS’s new 7.0 compiler), and only ONE that was just release for Linux, that will actually use this.
I wish I still had the URL, but in several CAD programs (which I would assume use similar operations as video games, please correct me if I’m wrong), Intel had mad a plug-in that re-compiles the program with vector support.
In these bench marks, with this compile, it made the AMD MP perform almost 60% faster in the top 6 operation (operations used the most). This was great! But the P4 results made my jaw hit the floor (Yes, I’m an AMD user, so don’t peg me as some Intel lover.). ALMOST 3 TIMES FASTER!!! It put the P4 (in the same top 6 operation) 1.4GHz ABOVE an AMD MP 1.2 (by a large margin, from being beaten into the ground without).
I was amazed, now, I know you would never see that kind of performance boost in real life, but still, when put in a native environment, the P4 SCREAMED.
I don’t expect to see a recompile of Direct X or Windows with this compiler (sense MS is now pushing there new compiler), but it’s already available for Linux. My question is, what happens in a year or two, when this compiler is available to game companies?
Remember, the P4 is in, what I would assume, a backward compatibility mode, and NOT it’s native processing mode with current compilers.
Also, I’ve heard that AMD is having yield problems past 1.4GHz. Is this true (I’d like an answer from someone that KNOWS, and not some know-it-all jerk please)? If this is true, what happens when the P4 goes to a smaller die and can hit even HIGHER clock speeds? Also, I’ve heard that more of the ORIGIAL design for the P4 (It was scaled back remember?) will be added back in. If this is true, does that not help even more so?
In the end, I half to ask, in two, three years, when we look back at the two chip, which one will end up being faster?
I’ve been out of the loop for a while, so go easy if my info is wrong; that’s why I’m here, to get it “updated”.
Peace!