Upgrading to 3gb Ram

[rldelrosario]

Good day again guys.

Here's my system:

OS: Windows ME/XP dual boot

Memory: 2 GiB RAM (2x 1 GiB)

Motherboard: P4X333-8235

CPU: Intel P4 2.0GHz

Video card: nVidia GeForce FX5500 256 MB (AGP 4x)

config.sys and autoexec.bat: no relevant entries

system.ini:

MaxPhysPage=48000 ; 1150 MiB

MaxFileCache=131072 ; 128 MiB

vmm32.vxd: plain vanilla (with 4.90.0.3000 vcache.vxd and vmm.vxd inside)

My question is this. Would it make a lot of difference if I add another 1gb ram. I usally use my pc to surf, office apps and nero burn.

Thanks for the help.

[bonestonne]

It could, but it wouldn't be a major difference. If you don't have trouble with 2gb now, you aren't going to see much of an improvement with 3gb.

I'm also not really sure about what you're doing that would need that extra 1gb, as 2gb of RAM seems like plenty for your usage and that system.

[Tripredacus]

You will probably only see an increase if you go to a higher memory speed, rather than size.

[puntoMX]

I would keep that system as is. If you need something that reacts and feels smoother you would be better off buying a new system.

[CoffeeFiend]

You will probably only see an increase if you go to a higher memory speed, rather than size.

Even that would hardly make any difference, about 1% or so for most applications. More RAM would likely give more of a boost if you use apps that need it (i.e. not just "surf, office apps and nero burn"), otherwise there is very little gain to be made either. Totally not worth the expense, when a modern $50 low end CPU (E3300; would cost about the same as new RAM) is about 550% faster, or something a little nicer, a $100 Athlon II X4 630 would be a bit over 1200% faster (not that they would fit on the old board admittedly).

Jus like puntoMX said, if it ain't fast enough then you're looking for a new computer (doesn't have to be expensive).

[dencorso]

My question is this. Would it make a lot of difference if I add another 1gb ram. I usally use my pc to surf, office apps and nero burn.

Not at all... at least no difference worth the effort and cost, IMHO.

But, just for Windows ME, you can tweak your settings up somewhat, say, to:

MaxPhysPage=7CB00 ; 1995 MiB

MaxFileCache=393216 ; 384 MiB

And safely find out whether more usable memory makes any difference for you, without needing either to buy anything, or to upgrade the hardware.

[rldelrosario]

good day guys...

thanks for all the replies.... guess the bottom line is that if i want something faster, might as well get a new cpu... might cost a little more but based on your suggestions, guess that would be a better deal.... for now, am cool with my unit, although an upgrade would definitely be much cooler...

again, thanks a lot mi amigos....

[Tripredacus]

You will probably only see an increase if you go to a higher memory speed, rather than size.

Even that would hardly make any difference, about 1% or so for most applications. More RAM would likely give more of a boost if you use apps that need it (i.e. not just "surf, office apps and nero burn"), otherwise there is very little gain to be made either. Totally not worth the expense, when a modern $50 low end CPU (E3300; would cost about the same as new RAM) is about 550% faster, or something a little nicer, a $100 Athlon II X4 630 would be a bit over 1200% faster (not that they would fit on the old board admittedly).

Jus like puntoMX said, if it ain't fast enough then you're looking for a new computer (doesn't have to be expensive).

I can tell the difference in memory speeds mostly in the current version of Flash used on websites. Otherwise, you are right, I wouldn't be able to tell very much.

[CoffeeFiend]

I can tell the difference in memory speeds mostly in the current version of Flash used on websites.

I have no idea how you'd even "evaluate" that in the first place (swap RAM for some faster clocked in the same computer, and then see if something in flash somehow works better?), and I don't see how memory speed would affect it in any perceptible way. It's usually either network bandwidth bottleneck (wait for it to download before it does anything), or otherwise a mostly CPU-bound process (i.e. decoding HD H.264 streams and the like). You got me curious on how you got to that conclusion.

[Zenskas]

Yeah memory speed is not that important, so long as it keeps your FSB happy you won't notice a difference between say 667 and 800MHz DDR2 (any 800MHz RAM will work with a 1600MHz FSB).

I have had my 800MHz RAM running at over 900MHz and saw a 0.5% increase on average in intense game benchmarks for starters. Doubt you would notice the increase of speed in web surfing and basic use.

[Tripredacus]

I can tell the difference in memory speeds mostly in the current version of Flash used on websites.

I have no idea how you'd even "evaluate" that in the first place (swap RAM for some faster clocked in the same computer, and then see if something in flash somehow works better?), and I don't see how memory speed would affect it in any perceptible way. It's usually either network bandwidth bottleneck (wait for it to download before it does anything), or otherwise a mostly CPU-bound process (i.e. decoding HD H.264 streams and the like). You got me curious on how you got to that conclusion.

Once a flash is loaded, it runs from the computer, not from the server. I'm not talking about watching videos like Youtube or JTV or something. Say we take a flash game (which I play on Facebook) and talk about the animation speeds. I've run 4 different configs in the past year and can see these differences in how Flash works. These results may seem strange but I can see the difference! Here is a list of configs in order of fastest to slowest:

1: Core 2 Duo 1.8GHz with 1.5GB DDR2@533MHz (7) 965 chipset

1: Pentium 4 2GHz with 512MB RDRAM@800MHz (xp) 850 chipset

3: Pentium 4 2GHz with 1.5GB DDR@333MHz (xp) 865 chipset

4: Pentium 4 2GHz with 1.5GB DDR@233MHz (xp) 865 chipset

Flash speeds between the RAMBUS and DDR2 is basically equal. Although the Core 2 seems to make up the difference between the memory speed with the RDRAM vs the P4 CPU without HT. I went from the RDRAM system to the DDR and it was horrible. I didn't have a choice because my board died. It was an upgrade on the Motherboard side, but a downgrade in the RAM. Going to 333 from 233 was a HUGE difference in speed on Flash. Not as fast as the RDRAM, as noted in my little chart.

[puntoMX]

You are talking about the badly designed P4 478, with a bus speed of 400MHz (quad pumped 100Mhz), so a 200MHz DDR (100MHz x2) stick would be in sync with the CPU, so, when you have a 800MHz RDRAM you would have twice the bandwidth over any DDR stick that can be useful, so, you are right with your statement there, however...

The 333 DDR and 266 DDR should not be much of a difference with those CPUs and I can hardly believe you can feel the difference; it must have been the timing on those DDR modules that gave you a smoother experience.

I don't know much about what flash does with memory or so, but it must be a lot of reads/writes at random addresses.

EDIT: The CPU on the i850, was that not a socket 423?

[CoffeeFiend]

I hate to say this but this sounds like the perfect example of how NOT to compare.

What you're saying is, based on your comparison across different machines with completely different CPU architectures, with a different amount of cores, varying clock speeds, different FSB speeds (and ratios), some with hypertreading and some without, using different sockets, wildly varying amounts of RAM, varying RAM types (3 kinds) *and* different clock speeds *and* different latency (and perhaps some with single or dual bank config, and with different interleaving), on different motherboards, with different chipsets (and most likely a whole lot more, including the video card) and the architecture changes that come with it (different interconnects & speeds), and running different operating systems that are most likely fairly different installs (in terms of configuration, drivers, background processes, etc), probably different browsers and/or flash plugin versions (build) -- and based on these totally different boxes you say that some gain in a completely subjective and unquantifiable speed evaluation taken across a fairly wide time range is due to RAM clock speed?

I mean, if you took the RAM from system 3 and put it in system 4 or vice-versa (and hand set the latency to the same on both -- don't rely on SPD) and somehow managed to do an actual measurement which showed repeatable difference beyond statistical irrelevance, then I'd have to say you must be right to some extent (it very well might still only hold true for older systems)

[Tripredacus]

I suppose you could be right, except for the Hyper Threading part. The P4 used in three of those examples was the same CPU, it did not support HT. I was going to buy one that did but then I came across the Core 2.