Learn how to a build a cloud-first strategyRegister Now

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 2605
  • Last Modified:

Memory bus width of a graphics card

Hello guys,

If a card has memory bus width of 128 bit, is it good, bad, or so-so? How does the width make any difference onthe performance of a card?

Thanks,
Yurich
0
Yurich
Asked:
Yurich
3 Solutions
 
jhanceCommented:
In general the wider the better since the onboard processor can move more bits of data into and out of the onboard memory with each bus transaction.  These days 128-bit video cards are mainstream and I would say good.  In evaluating the performance, however, there are MANY factors other than simply memory bus width that need to be taken into account.
0
 
Purple_SkyCommented:
128 mb is acceptable. Clock rates are very important too.

Modern 3D graphics cards rely on the values of many variables in order to determine their speed. The most important of these variables are; clock speed, memory speed, and pixel processing volume. One very important card specification that is often overlooked is the number of pixel pipelines. If a card has 8 pixel pipelines, a clock speed of 400Mhz, and a memory clock of 1000Mhz, it is often slower than a card with 16 pipelines and lower clock and memory clock values.

0
 
garycaseCommented:
Remember that a modern graphics card is really just another computer with a very specialized command set.   The CPU sends graphics commands to it, and it executes those commands.

So ... as with any computer, the wider the path to memory, the more data you can transfer between the GPU and memory on each access.  The memory clock speed determines how quickly those accesses occur -- so as noted above it's not a simple "just look at one thing" equation.

The reason it's so difficult to compare the absolute performance of graphics cards is the large number of variables involved:  the speed of the GPU (and even the same basic GPU is usually available in many different variants), the width of the path to memory (per this question), the speed of the memory, how much memory they have, the clock speed of the GPU, the number of pixel pipelines (paths to the actual graphics translator), the interface to the primary system's memory (AGP, PCI, PCI-X, PCI Express, and the various variants of those ==> AGP 2x, 4x, 8x; PCIe x16, x4, x8, x1, etc.), etc.

Bottom line:  "... is it good, bad, or so-so ..."  ==> of those choices, I'd say it's "so-so".   It's simply one factor (of many) that determines the overall performance of the graphics processor.   The current top-of-the-line cards all have 256-bit memory interfaces and 512mb of memory; but that doesn't in any way mean that a narrower memory interface or less memory is "bad."   Those top-of-the-line cards also have VERY fast GPU's that can utilize the faster interface and additional memory.  They also have commensurate price tags :-)
0
 
YurichAuthor Commented:
Thank you guys,
Yurich
0
 
garycaseCommented:
You're most welcome.
0

Featured Post

Independent Software Vendors: We Want Your Opinion

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

Tackle projects and never again get stuck behind a technical roadblock.
Join Now