Right, I am saying that without a doubt *some* game reviews are skewed higher by the influence of PR guys working for the developer or publisher. They are not always the untainted opinion of the reviewer(s).
I've been in the games industry for a while now and can tell you any one familiar with the inner workings of Game PR and Marketing would not be surprised by this, or the allegations being made. Many things happen between game media and game developers / publishers that result in a less than objective review. I won't spill all the beans, but for example: You know those top ten lists? You can pay to have your game show up on most of those. Companies I've worked at have bartered exclusive first looks at new titles for higher review scores on shipping titles. Our PR guys even wrote up material for reviewers to use in their reviews that highlight features we wanted to the public to get excited about. Those reviews are a lot less about journalism and are really an extension of Studio/Publisher PR and Marketing machines. Don't buy what any of the big sites say, just download the demo and make your own call.
The hardware thread thing on the Xbox360 is kind of important. Each CPU issues 2 instruction every cycle and as long as they don't all end up needing to go to the same execution unit, they will be processed at that rate.
It should be pointed out that on most modern CPU's you average probably ~30% CPU utilization and the rest of the time is spent for D and I cache to fill. Any one who has programmed on PS2 and seen a T15k profile can back me up on this. I think V-Tune PC results are similar as well. Because of the way the Xbox360 CPU will flush a thread while its stalled waiting for cache to fill it will always be trying to run something, so in theory the CPU utilization will be better. This is the opposite of the way the PS2 worked. It could execute two instructions in the same cycle, but they had to be from the same thread and meet certain criteria. The end result was this other execution pipe seeing ~10% utilization average across a frame. Although, on Xbox360 since threads share cache mismatching threads with different purposes on the same CPU at the same time will cause more than average cache thrashing.
The playstation3 SPE's don't have cache in the tradition sense we think of L1 or L2 cache. Its got a block 256kb of memory we more or less explicitly fill. This filling is more expensive than traditionally filling L1 or L2 BUT forces us to pretty much write code that works within those bounds. The result is code that effectively never has cache misses so CPU utilization should be very very high. Pretty much they traded small high frequency memory transactions for large very low frequency ones. I expect the results on final hardware to be pretty spectacular.
The Xbox360 CPU's are general purpose, and have the full PPC and Altivec units for each core.(well, almost, some altivec instructions are missing and many more were added, but this is just for x360). The Playstation3 on the otherhand has 1 general purpose PPC CPU and 6 to 8 cell processors, which are SIMD vector (but not altivec) units and not general purpose in the least.
The Xbox360 is quite capable of running any C++ game logic, while the SIMD nature of the playstation3 cells make C++ not a good choice at all. Your original assesment about it being only a good number crunching machine is not quite right for x360 but totally fits the Playstation3.
I'm honestly not sure how well this processor would work in a more general purpose PC setting though, mainly due to the fact that each core is running two threads (also issues 2 instructions per clock) so cache space is a bit of a concern. In games its easy to put 2 like threads on a CPU because we have very explicit control over it. On a PC this isn't really the case, and massive cache thrashing would be had.
RIMMER: (VO) After intensive investigation, comma, of the markings on the alien pod, comma, it has become clear, comma, to me, comma, that we are dealing, comma, with a species of awesome intellect, colon.
HOLLY: Good. Perhaps they might be able to give you a hand with your punctuation.
Re:Climbing(i apoligize for not explaining jargon)
on
Sports Technology?
·
· Score: 1
No, not really.
Even 25 years ago hard stuff was being sent, i think the FA on Midnight Lightning (V8/9) came in the late 70's or early 80's when the rest of climbing was stuck in the 5.11 and 5.12 range. The think that held roped climbing back was far from technology. To sport climbing as we know it, there has not been any advance in technology in 50 years that has advanced the sport to unknown limits.
The thing that allowed climbing to progress more than anything was a change in ethics. Back in the 70's and 80's it was considered cheating to use bolts and draws as a form of protection, so a lot of hard routes were done trad style. The primary reason was the unsighlyness of the bolts, but also because some people have trouble letting go. Also the way we work routes, hanging at each move we can't get until we figure it out was considered cheating too. People used to try a route once and if they fell, lowered all of the way to the ground to try it again. At this point in time that was the only way to get a true 'redpoint', hanging until you could get the move earned you the label a 'hangdog' and any send you eventually could get was called a 'pinkpoint'. Chalk wasn't widely used until the 80's because it was considered to have unsightly consequences just like bolting, and is still banned in some places (skeleton cave, OR for example)
Bouldering however, the moves were still as hard, but there were no silly things holding people back from thier potential. Since climbing became what it is now, it has been advanced by an elite cadre of freaks of nature who can do one arm pull-ups off of thier pinky fingures. **cough**fred**cough** and it'd be silly to think that anyone could climb that hard.
I'm really curious how much of that contract is for actual product and how much is for consulting and maintenance for actually building the system.
I used to work for a company that made business software and would often sign deals with organizations like the Army and Navy for multiple thousands of seats of our product. A good part of the contract was for all the consulting and maintenance we'd provide to them to get it all set up and any further help they would need for the next year or so. Its not a bad model, because after their initial year of maintenance ran out, we'd get to renew that part of the contract for another large sum of money and repeat.
Slashdot Mirror
Right, I am saying that without a doubt *some* game reviews are skewed higher by the influence of PR guys working for the developer or publisher. They are not always the untainted opinion of the reviewer(s).
I've been in the games industry for a while now and can tell you any one familiar with the inner workings of Game PR and Marketing would not be surprised by this, or the allegations being made. Many things happen between game media and game developers / publishers that result in a less than objective review. I won't spill all the beans, but for example: You know those top ten lists? You can pay to have your game show up on most of those. Companies I've worked at have bartered exclusive first looks at new titles for higher review scores on shipping titles. Our PR guys even wrote up material for reviewers to use in their reviews that highlight features we wanted to the public to get excited about. Those reviews are a lot less about journalism and are really an extension of Studio/Publisher PR and Marketing machines. Don't buy what any of the big sites say, just download the demo and make your own call.
The hardware thread thing on the Xbox360 is kind of important. Each CPU issues 2 instruction every cycle and as long as they don't all end up needing to go to the same execution unit, they will be processed at that rate.
It should be pointed out that on most modern CPU's you average probably ~30% CPU utilization and the rest of the time is spent for D and I cache to fill. Any one who has programmed on PS2 and seen a T15k profile can back me up on this. I think V-Tune PC results are similar as well. Because of the way the Xbox360 CPU will flush a thread while its stalled waiting for cache to fill it will always be trying to run something, so in theory the CPU utilization will be better. This is the opposite of the way the PS2 worked. It could execute two instructions in the same cycle, but they had to be from the same thread and meet certain criteria. The end result was this other execution pipe seeing ~10% utilization average across a frame. Although, on Xbox360 since threads share cache mismatching threads with different purposes on the same CPU at the same time will cause more than average cache thrashing.
The playstation3 SPE's don't have cache in the tradition sense we think of L1 or L2 cache. Its got a block 256kb of memory we more or less explicitly fill. This filling is more expensive than traditionally filling L1 or L2 BUT forces us to pretty much write code that works within those bounds. The result is code that effectively never has cache misses so CPU utilization should be very very high. Pretty much they traded small high frequency memory transactions for large very low frequency ones. I expect the results on final hardware to be pretty spectacular.
Not quite.
The Xbox360 CPU's are general purpose, and have the full PPC and Altivec units for each core.(well, almost, some altivec instructions are missing and many more were added, but this is just for x360). The Playstation3 on the otherhand has 1 general purpose PPC CPU and 6 to 8 cell processors, which are SIMD vector (but not altivec) units and not general purpose in the least.
The Xbox360 is quite capable of running any C++ game logic, while the SIMD nature of the playstation3 cells make C++ not a good choice at all. Your original assesment about it being only a good number crunching machine is not quite right for x360 but totally fits the Playstation3.
I'm honestly not sure how well this processor would work in a more general purpose PC setting though, mainly due to the fact that each core is running two threads (also issues 2 instructions per clock) so cache space is a bit of a concern. In games its easy to put 2 like threads on a CPU because we have very explicit control over it. On a PC this isn't really the case, and massive cache thrashing would be had.
And maybe i can figure out that whole newline thing.. i kick ass in my hypocrisy.
RIMMER: (VO) After intensive investigation, comma, of the markings on the alien pod, comma, it has become clear, comma, to me, comma, that we are dealing, comma, with a species of awesome intellect, colon. HOLLY: Good. Perhaps they might be able to give you a hand with your punctuation.
No, not really.
Even 25 years ago hard stuff was being sent, i think the FA on Midnight Lightning (V8/9) came in the late 70's or early 80's when the rest of climbing was stuck in the 5.11 and 5.12 range. The think that held roped climbing back was far from technology. To sport climbing as we know it, there has not been any advance in technology in 50 years that has advanced the sport to unknown limits.
The thing that allowed climbing to progress more than anything was a change in ethics. Back in the 70's and 80's it was considered cheating to use bolts and draws as a form of protection, so a lot of hard routes were done trad style. The primary reason was the unsighlyness of the bolts, but also because some people have trouble letting go. Also the way we work routes, hanging at each move we can't get until we figure it out was considered cheating too. People used to try a route once and if they fell, lowered all of the way to the ground to try it again. At this point in time that was the only way to get a true 'redpoint', hanging until you could get the move earned you the label a 'hangdog' and any send you eventually could get was called a 'pinkpoint'. Chalk wasn't widely used until the 80's because it was considered to have unsightly consequences just like bolting, and is still banned in some places (skeleton cave, OR for example)
Bouldering however, the moves were still as hard, but there were no silly things holding people back from thier potential. Since climbing became what it is now, it has been advanced by an elite cadre of freaks of nature who can do one arm pull-ups off of thier pinky fingures. **cough**fred**cough** and it'd be silly to think that anyone could climb that hard.
I'm really curious how much of that contract is for actual product and how much is for consulting and maintenance for actually building the system.
I used to work for a company that made business software and would often sign deals with organizations like the Army and Navy for multiple thousands of seats of our product. A good part of the contract was for all the consulting and maintenance we'd provide to them to get it all set up and any further help they would need for the next year or so. Its not a bad model, because after their initial year of maintenance ran out, we'd get to renew that part of the contract for another large sum of money and repeat.