When shopping for a video card, there is no doubt that a prime feature bragged about the card is how much memory it has and what kind. But, what does that mean? And how does the type and amount of memory effect the performance of the video on your system? Well, read on…
Simply put, a video card relies on its memory to draw a picture. Different cards offer different amounts of this precious memory. While 4 to 8 MB used to be pretty good, now days, seeing 32 MB up to 128 MB of memory on a video card doesn’t raise anyone’s eyebrow.
Calculating Memory Usage
To put it simply, the more memory, the better. For 2D applications, it doesn’t really matter. But, for 3D games, it is optimum to have gobs of memory on your video card in order to store the textures the game generates and thus be able to deliver the pictures quickly.
If you don’t play games, there is a mathematical equation to help you figure out just how much video memory you need. This depends on your resolution. The higher the resolution, the more memory you need, because each and every pixel on your screen must have a space in the memory for its data. Here is the computation: Lets say you would like to display 256 colors on a screen resolution of 640×480. At this resolution, there is 307,200 dots, or pixels. 256 colors requires 8 bits or data for each pixel. You can figure this because with an eight digit binary, there are 256 possible combinations. For two colors, you need only 1 bit, either on or off. For 16 colors, you need 4 bits, 2 to the 4th power. 256 colors requires 8 bits, and it goes up from there. Anyway, multiply the number of dots by the number of bits per pixel to get the number of bits for the entire screen.
307,000 x 8 = 2,457,600 bits.
There are eight bits per byte and 1,024 bytes per kilobyte. So…
2,457,600 / 8 = 307,200 bytes = 300K
Therefore it requires exactly 300K of memory to display 256 colors at 640×480 resolution. But, after calculating this, you must consider the available amounts. You cannot buy a video card with 300K of memory. They were available at either 256K or 512K. So, to get this resolution and color scheme, you must buy a card with 512K of memory on-board.
Obviously, though, most users do not deal with low memory amounts such as this anymore. But, most users do not use low 256 color levels anymore, most opting for at least 16-bit color. So, higher amounts of memory are needed even for standard 2D business software, although it is nothing compared to the requirements of full motion 3D. Of course, making the memory faster also helps overall speed, which is why newer kinds of video RAM were developed.
The gobs of memory available on card today are there for two reasons: (1) a marketing ploy: it is our nature to think that the more memory we have, the better and we want beefed up systems that will arouse a nerd. (2) Gaming: most cards on the market today are designed primarily for 3D gaming applications, with 2D being almost an afterthought. With games becoming more life-like, pumping more polygons and using richer textures which all need to be stored in memory, there definitely is a need for a lot of video memory. AGP video cards have the added ability to be able to also use the main system memory as part of the overall memory pool, so the video card’s own memory is no longer the cut-and-dry limitation it once was. That said, video RAM is a special kind of RAM designed with low latencies needed for high speed video environments, so whenever the video card can use its own memory over the system memory, performance benefits.
Another issue with regards to memory is the local bus between the video chipset and the video memory. Basically, the wider the access to memory from the chipset, the faster the card. This access is a local bus wired into the card. When shopping, you will often see 64-bit or even 128-bit cards advertised. This refers to this memory bus. It is often confused with the bus slot the card plugs into. In reality, a 64-bit card means that it has a 64-bit memory bus on the card, but is actually a 32-bit card plugging into a PCI bus slot.
Types of Video RAM
Most past video cards used a type of video memory called Dynamic RAM, or DRAM, to store image information. This type of RAM was easy to make and was, therefore, cheap. The problem was that it was too slow. This slowness stems from the fact that DRAM needs to be constantly refreshed in order to save the info. DRAM also cannot be read at the same time as it is being written to. This slowness could be demonstrated by looking at one of these cards. At a standard resolution of 1024×768 with a 72Hz refresh rate, the DAC (Digital to Analog Converter) needs to access the memory’s information 72 times per second. If using 24-bit color, this translates to about 170M/sec of transfer from the DRAM, the max transfer of DRAM. This does not include the writing of new information, which can’t be done during the read process.
In response to this, a number of new memory technologies have been developed:
- EDO RAM was the first to be used. EDO RAM was often used in the main system memory as well. EDO RAM, when used for video, was only slightly faster than DRAM, but the cost was about the same. It worked by offloading its contents to separate circuits, thereby allowing it to receive new data at the same time that the last data cycle is being completed. EDO RAM is architecturally the same. The only difference was with the wiring onto the board.
- Video RAM, or VRAM, was used on video cards for awhile. VRAM is dual-ported, meaning it allows all hardware to access the memory at the same time, including the main processor, the video processor, the DAC, etc. It was more expensive than EDO, but the performance increase was noticeable.
- Window RAM, or WRAM, is modified VRAM. It boasted slightly better performance than VRAM and cost less to make. For this reason, it was often used in place of VRAM on many video cards.
- MultiBank DRAM (MDRAM) was aimed at the cost-sensitive user who requires good performance. The downside is that it is rather rare. The prime difference is its organization. While standard memory was limited to sizes or 256K, 512K, 1M, 2M, 4M, etc, MDRAM could be added in increments of 32K. This allows the user to calculate just how much memory they really need, using the above equation, and upgrade their video memory to be very close. Performance wise, MDRAM was much faster than VRAM or WRAM.
- Synchronous Graphics RAM, or SGRAM, and SDRAM are used on many modern cards today where speed is an issue. SGRAM is clock-synchronized, like SDRAM used for main system memory. It was relatively low-cost, but had a few features that made it a good performance video memory. It used masked write, which enabled certain data to be modified in a single operation rather as a sequence of read, update, and write operations. It also used block write, which allowed data for background or foreground image fills to be handled more efficiently. SGRAM was single-ported, but it could access two memory pages at once, thereby simulating the dual-port of VRAM and WRAM.
- Double Data Rate (DDR) SDRAM is the type of memory used mostly today. It is the same as that used for main system memory.
The PCMech.com weekly newsletter has been running strong for over 8 years. Sign up to get tech news, updates and exclusive content - right in your inbox. Also get (several) free gifts.