Benchmark Results: Synthetics
With so little performance difference in games and applications, perhaps synthetic benchmarks will show a larger variation in potential performance?
The Sandra Memory Bandwidth benchmark improves when more banks are filled, yet other synthetics do not appear to get a noticeable benefit from the added bandwidth.
232
Comments
Related Stories :
- Memory Module Makers See Strong July Sales
- Memory Module Makers Sharply Reduce NAND Flash Inventory Amid Price Slump
- Thermaltake launches Cyclo module memory cooler
- NAND flash oversupply provides memory module makers more freedom in platform adoption
- Memory module makers pushing NAND flash devices at Computex 2005
- Memory-module makers could face a tough 2005
- Memory Module with MP3 Player: MuVo NX
- Team Group Announces First 3GHz DDR3 Memory Modules
- AMD Intros Branded Desktop Memory Modules
- Corsair Announces 8GB DDR3 Memory Modules
See more
Sponsored
All about Memory
I run 8GB in my rig because... having open memory slots doesn't jive well with my OCD. Haha. /facepalm
Try opening a PSD file from a 21-Megapixel camera with 3 or more layers in Photoshop and see the difference.
Great info for the Vista crowd. What about DDR2 and XP? Betcha there are more of us than youse guys.
can someone help me with what apps will utilise more ram?
im currently researching building a workstation for photoshop and some 3d modelling to be imported and rendered in CS4
any help much appreciated im no hardware guru
It's generally accepted that XP uses less RAM than Vista, so you should be fine with 4GB (which, with 32-bit XP, will yield around 3GB of usefull address space).
@Thomas Soderstrom
What the heck is that image on the front with all the memory modules!
Can you supply a link with a bigger pic??? That's just insane!
Hey guys I was wondering: What if you had a 9g game loaded completely on a ramdrive (64bit+software) then kept the remaining 3 for regular ram?
I got 12 for all your application pleasure baby *wink*
@Thomas Soderstrom What the heck is that image on the front with all the memory modules!Can you supply a link with a bigger pic??? That's just insane!
It looks like a memory testing machine or burn-in rack to me...I'm not sure where the site found it.
Hey guys I was wondering: What if you had a 9g game loaded completely on a ramdrive (64bit+software) then kept the remaining 3 for regular ram?
That's a great idea! Because it takes sooo long to copy an entire game from the hard drive to "anything" (even RAM), most users won't even try it. But you would get super-fast map loads!
Try using several VMs in parallel and you'll be happy to have more than 3 GB. I'm using one VM for work (image of my work notebook with VPN etc.) and one for online banking (mainly because my old smartcard isn't supported in 64bit OS, little bit of a catch22).
Because it takes sooo long to copy an entire game from the hard drive to "anything" (even RAM), most users won't even try it. But you would get super-fast map loads!
I have 8GB and use 4GB to load a MMORPG game on and the other 4 to use for system ram. It's great to cut down random disk access and helps improve load times. Like you mentioned though the downside is it takes awhile to copy your game data or whatever into the ram drive. My ramdrive is setup to copy the data automatically on startup, but startup time takes a modest amount longer due to that.
What about setting up RAMdisks (as a temp folder, working directory, swap partition, etc)... surely doing this with 6-12GB of RAM would make a big difference?
This article illustrates the dark side of Tom's Hardware: making a big deal about quantifying performance differences that are of no practical impact. In this case, even the concept for the article is idiotic. Why would you even THINK to look for differences in application performance beyond the threshold of adequate memory? There's a bit of value in the last couple of pages, but beyond that you've wasted your time and ours.
I'm not saying this just for the entertainment value of trolling. I want you to know that publishing crap like this discredits the Tom's Hardware brand. How many articles like this will people read before writing off Tom's Hardware completely as a source of information?
I've got 12GB of OCZ's DDR3-1600 overclocked to 1780 in my system. With triple channel 6GB kits at under $100, there's no reason not to populate all the 6 (or 4) DIMMs on a motherboard.
Also, if you've ever a virtual OS the windfall of having over 3Gb is huge.
Having more memory is about system response to load times and clealy multitasking, try to open Premiere , Photoshop and After Effects at the same time and do your work and you will see a huge impact into system response not just stupid 1 task gameing. Just 2 examples:
http://img5.imageshack.us/img5/1300/30630315.jpg
http://img5.imageshack.us/img5/1691/40873236.jpg
I guess this would run just fine on 3 GB too ?
After reading this I wish I would have never bought that second kit of memory. Nothing but a waste of money.
The conclusion says: "We can only recommend larger capacities of 8 GB to 12 GB for professional applications where its usefulness has already been documented and for servers."
Therefore it is meaningless to mention about the memory requirement for 3 Adobe programs at the same time. Just read the article carefully.
Having more memory is about system response to load times and clealy multitasking, try to open Premiere , Photoshop and After Effects at the same time and do your work and you will see a huge impact into system response not just stupid 1 task gameing. Just 2 examples:http://img5.imageshack.us/img5/1300/30630315.jpghttp://img5.imageshack.us/img5/1691/40873236.jpgI guess this would run just fine on 3 GB too ?
Just read the end of the article carefully.
"We can only recommend larger capacities of 8 GB to 12 GB for professional applications where its usefulness has already been documented and for servers"