palladin9479 :
tourist :
palladin9479 :
tourist :
Gulli :
^^
So basically you are saying overclocking RAM by a third is reliable and easier than just closing browser tabs before you start a game (and that's assuming everyone just has 20 browser tabs opened all the time, something that wouldn't be fixed with more memory because someone who opens 20 tabs does so because he can and will just open 40 tabs if he has more memory)?
P.S. where I live you can buy 4GB of faster RAM for the same price as 8GB of slower RAM. Maybe there's no rebate action but rebates are misleading in the first place for anyone reading this article more than a day after it was first published. Besides I'm annoyed how easily fictional RAM requirements (how much people think they need) double every couple of years without actual games and applications coming close to utilizing all that RAM.
Your missing the point, when the apu reserve up to two gb of ram that only leaves 1.5 gb for windows use. Are you saying 1.5 gb is enough to play modern games with windows 7/8 ?
Swapping files in and out from the hd even a ssd causes slowdowns. This is because windows dynamic memory allocation knows you only have 4gb and will make due with what it can ram,or hdd. This is not a dedicated gaming rig where you turn it on and play a game..
That's not how memory works with an APU. The dedicated chunk gets taken out ~before~ the 32 bit memory cap of 4GB is imposed.
Is that [strike]not what i said above[/strike]Also you can chose how much memory to allocate to the iGPU and picking 2GB would generally be a bad idea as you won't be running high AA and other memory intensive settings anyway. 1GB is usually the best setting.
So 1GB for iGPU and 3GB for the OS. 32-bit NT applications can't use more then 2GB (31 bit address space) anyway so unless your also running 64-bit applications your already at your absolute memory limit. with 4gigs yes you are.
But windows memory allocation will go higher with 64bit 8gigs.
All in all 4GB of ultra fast memory will give you better performance then 8GB of slower memory and since we're talking an ultra tight budget it might be worthwhile to investigate that route.
I tried to answer this on my s2above, but it was a cluster.
win 32 will
address up to 3.5 gb of memory for the whole system. It matters not if you use 4gig or 8gigs. The only thing the bios does is
RESERVE xxx memory in windows to run graphics aplications, aero, , flash, games etc. It does not mean windows will not allocate more but will always reserve a minimum of xxx for graphics as a priority. with 4gb you are forcing windows to swap back and forth to retain sufficient headroom to run background programs while playing games as up to 2gb will be reserved for graphics. 1 gb is ok but in this world users see 2gb and set it to that.
Umm that's not how the NT memory model works. NT (MS Windows) use's 32 bits to address all system memory for a total of 4GB of address space. That space is then divided into two segments of 31 bits each with the last bit being used as a separator. The lowest 31 bits are for applications address space with the upper 31 bits being used for the kernel address space. It's a form of security that dates back to NT 4.0 and is required to maintain backwards compatibility at the binary level. The application space is a virtual address space and you can have as many of them as you have paging space for, actual memory pages are only allocated when needed. So having five applications running, each will have 31 bits of virtual address space (2GB). The NT Kernel is a monolithic kernel and all kernel components, drivers and libraries share the same 31-bit address space of 2GB.
Now to talk to anything hardware related you need a memory address. The IBM PC standard has set locations for some things under 1MB (VGA / BIOS / ect..), Plug and Play remaps everything into the last 512MB ~ 1024GB address range. Back then systems had 16~32MB of memory and the idea of a client having 4GB of RAM was ludicrous. So now you have the BIOS, VGA BIOS, and any special motherboard related components all mapped into the last 512MB of space. Seriously look at device manage resources and you can see the mappings. MS Windows has absolutely no problem actually using 4GB of memory, the last portion of the
address space is being used which prevents the OS for directly addressing system memory that would otherwise be there.
Now how does this deal with iGPU memory? In the past it was directly mapped into the upper range and could easily take 256~512MB of memory. You can see this on certain old MB's that only show 3GB of system memory. Once graphics cards passed the 1GB memory they adopted a paging system that would only map a portion of the VGA memory into the virtual address space and the drivers page it in and out in a style similar to EMS. It's a work around for the older NT x86 memory model.
So using the 32-bit version of Windows XP on this system would give you 4GB of memory with the first 1GB being hard mapped (the OS can't change it) to the iGPU and Windows using the remaining 3GB for itself. That 3GB is all dynamically allocated and the kernel tends to only use 512MB of space with the remaining 2~2.5 GB being available to all applications on the system.
Now the ~real~ answer to this problem is to stop using f*cking Windows XP. The 64-bit NT memory model does away with all that nonsense.
Really pallandin, ? you want to get this far into the weeds for something so simple. Run the test yourself with 4gb in win 7 32bit start up a skyrim or even wow and watch you swap file grow
Second your understanding of uma graphics memory is outdated
From Tech power up:
The integrated graphics of all AMD A-Series uses the system's main memory as graphics memory. There are two ways how (system) memory can be allocated as graphics memory: UMA (Unified Memory Architecture) and AMD HyperMemory.
The allocation size for UMA is set in the BIOS and dedicates that amount of memory to the GPU, away from the CPU and Windows will have no access to this memory slice at all. Once UMA memory runs out, AMD's driver has a feature called HyperMemory which lets the driver allocate additional memory from within Windows, like any other application would do, to be used as graphics memory.
UMA has the advantage that the graphics hardware has access to a single, dedicated, unfragmented memory range that it can work its magic in without worrying about the rest of the system. Here, the disadvantage is that the memory is assigned to the graphics chip, even when the graphics hardware does not need that much memory, for example when only the Windows Desktop or 2D applications are active.
HyperMemory solves the problem of unused memory by dynamically allocating as much memory as the GPU needs, when it needs. As a result more (system) memory is available for normal applications to use, but performance of the graphics memory is reduced. Since the driver has to manage this memory allocation, which is in turn passed on to the Windows Kernel, several layers of processing are added which results in a significant performance impact as we see in the following results.