CS5 performance question

[zack10]

Colin,
If it's "The Plague", why does my computer buzz through tasks like never before? I guess HP's tech guys all are wet. I have 8GB and 83%(6GB) set for CS5. Works way good Mon!

GT

[Colin Southern]

If it's "The Plague", why does my computer buzz through tasks like never before?

It should buzz through them anyway. Some things to keep in mind ...

- Writing to / reading from RAM is much much much faster than writing to / reading from a HDD, therefore you want to PC to be using RAM for it's memory, not the hard drive.

- For a given set of programs & their data requirements, the ONLY way to do that is to have sufficient RAM so that the PC doesn't need to page to HDD

- The OS default max page size is set to 3x physical RAM size - which is what the system would have used on automatic settings anyway.

- If you leave it set on automatic then it'll grow to that size if it needs to - so not a lot of point in setting a minimum size (especially when it's desireable not to use paging anyway).

I guess HP's tech guys all are wet.

Not at all - very clued up bunch (I WAS the sole HP tech for Pavillion PCs, large format printers, and laser printers for our city for many years, and an HP beta tester)

I have 8GB and 83%(6GB) set for CS5. Works way good Mon!

I have 4GB with 100% set for CS5 at home, and 12GB with 100% set for CS5 at work (on an i7 Extreme Edition CPU, dual 8800GTX graphics cards (in SLI config), Dual Western Digital Raptor drives in RAID 0 for OS, and another pair of drives in RADI 1 for data) and mine both work way good too mon

[carregwen]

... and 12GB with 100% set for CS5 at work (on an i7 Extreme Edition CPU, dual 8800GTX graphics cards (in SLI config), Dual Western Digital Raptor drives in RAID 0 for OS, and another pair of drives in RADI 1 for data) and mine both work way good too mon

I'd want my money back if it didn't

[proseak]

It's worth remembering how Photoshop uses virtual memory - it doesn't share the pagefile, but rather uses it's own, dedicated system known as scratch discs. You'll find access to them in the "Performance" tab. It is important to set them on a disc other than the system disc(or wherever you keep your pagefile) so if sys is "C" and you have C, D, and E to allocate them D, E, C. (or E, D, C). This becomes really important with larger files, and longer edit-history.

Some weeks ago I had a power outage so Photoshop couldn't clean up after itself. The size of the scratch file was 55Gb!

[zack10]

It should buzz through them anyway. Some things to keep in mind ...

- Writing to / reading from RAM is much much much faster than writing to / reading from a HDD, therefore you want to PC to be using RAM for it's memory, not the hard drive.

- For a given set of programs & their data requirements, the ONLY way to do that is to have sufficient RAM so that the PC doesn't need to page to HDD

- The OS default max page size is set to 3x physical RAM size - which is what the system would have used on automatic settings anyway.

- If you leave it set on automatic then it'll grow to that size if it needs to - so not a lot of point in setting a minimum size (especially when it's desireable not to use paging anyway).



Not at all - very clued up bunch (I WAS the sole HP tech for Pavillion PCs, large format printers, and laser printers for our city for many years, and an HP beta tester)




I have 4GB with 100% set for CS5 at home, and 12GB with 100% set for CS5 at work (on an i7 Extreme Edition CPU, dual 8800GTX graphics cards (in SLI config), Dual Western Digital Raptor drives in RAID 0 for OS, and another pair of drives in RADI 1 for data) and mine both work way good too mon

Colin,
Now that you cleared this up, why would HP have me do this in the first place? You also said HD space is wasted, in my case 16GB, but the free space value for drive "C' hasn't changed. How can this be?
GT

[Colin Southern]

why would HP have me do this in the first place?

Lets look at the question from a different angle; if such performance gains were to be had, why wouldn't Microsoft just do it this way in the first place? Big performance gains - no apparent overhead - you'd think it would be a no-brainer.

You also said HD space is wasted, in my case 16GB, but the free space value for drive "C' hasn't changed. How can this be?

Probably because the pagefile had already been created.

Honestly, I've been dealing with PC performance issues from the time the IBM PC was invented in the early 80s, (and yes I had one - complete with 16KB RAM - a 360k 5.25" FDD - and a cassette port - HDDs - the first 10MB HDD hadn't been invented yet) (in fact, I've made a career out of it) - I've upgraded hundreds of PCs suffering from thrashing (using the HDD as additional RAM heavily) - trust me, using the swap file may allow a program to continue (I've done it myself when combining over a dozen images at the same time on a laptop), but paging to disk in lieu of sufficient RAM NEVER increases the speed of the PC - it slows it down to a crawl because the bottom line is RAM is a lot faster than a hard drive. It can buy a bit of breathing space if you have a lot of things open, but not actually doing anything (the kernel memory management will swap out pages of RAM that aren't currently being used often to give priority to those that are needed more often), but the best solution of all is still to have enough RAM to keep the PC happy in the first place. When the going gets tough, a PC with 24GB RAM and no pagefile will beat the living daylights out of one that only has 6GB RAM and an 18GB page file.

Hope this helps; and on that note, I'm going to move on. Pickup a copy of Windows System Internals by Mark Russinovich & David Solomon 5th edition if you want to learn more than you ever wanted to learn about the topic

[zack10]

Lets look at the question from a different angle; if such performance gains were to be had, why wouldn't Microsoft just do it this way in the first place? Big performance gains - no apparent overhead - you'd think it would be a no-brainer.



Probably because the pagefile had already been created.

Honestly, I've been dealing with PC performance issues from the time the IBM PC was invented in the early 80s, (and yes I had one - complete with 16KB RAM - a 360k 5.25" FDD - and a cassette port - HDDs - the first 10MB HDD hadn't been invented yet) (in fact, I've made a career out of it) - I've upgraded hundreds of PCs suffering from thrashing (using the HDD as additional RAM heavily) - trust me, using the swap file may allow a program to continue (I've done it myself when combining over a dozen images at the same time on a laptop), but paging to disk in lieu of sufficient RAM NEVER increases the speed of the PC - it slows it down to a crawl because the bottom line is RAM is a lot faster than a hard drive. It can buy a bit of breathing space if you have a lot of things open, but not actually doing anything (the kernel memory management will swap out pages of RAM that aren't currently being used often to give priority to those that are needed more often), but the best solution of all is still to have enough RAM to keep the PC happy in the first place. When the going gets tough, a PC with 24GB RAM and no pagefile will beat the living daylights out of one that only has 6GB RAM and an 18GB page file.

Hope this helps; and on that note, I'm going to move on. Pickup a copy of Windows System Internals by Mark Russinovich & David Solomon 5th edition if you want to learn more than you ever wanted to learn about the topic

Colin,
Thanks for taking the time to explain this.
GT