Introduction
I am pleased to see the keen interest that many of you have in AGP related graphics performance. In truth, the issue is multidimensional. In this article, I will expand on more AGP topics, but let me warn everybody that the bottom line analysis is still the same.
Before I get into the details, allow me respond to a few of the straightforward questions that have come up.
Is PCI better than AGP?
No. AGP is more powerful, more flexible, and has more potential than PCI. If you have the choice, and everything else is equal, AGP should be preferred. However, the power user must be made aware that this flexibility can be used to create resource conflicts that might be unfavorable for ultimate game performance. In contrast, these trade-offs may be perfectly acceptable for the mainstream user. On the whole, the urgency to invest in an upgrade to AGP is not all that it is cracked up to be.
Is the Real3D Starfighter PCI a good product.
I have not completed all of my testing on it, but I believe it to be an excellent product, and I believe their AGP product to be an excellent product (with potential trade-offs that should be understood). The PCI product has a generous 8 or 16 megabytes of dedicated texture memory that will prevent the main memory bus conflict that might occur with an AGP card. Above all, let me add that the crew at Real3D are extremely talented and a pleasure to work with.
Is AGP a Performance Risk?
The issue here is texture management in cases where there is insufficient graphics memory. AGP provides a superior alternative to PCI in dealing with texture management problems under these cases. I still do not believe that the AGP solution is better than simply ensuring that you have an adequate graphics memory configuration.
Don’t I need AGP2x or 4x for the future?
All AGP4x does is allow the graphics controller to assert more of its texture bandwidth demand on main memory. This is the source of the performance conflict, and 4x allows it to become a bigger problem. But there are two sides to examine: Texture demand and geometry transfer rates. We will do that in a minute.
When it is all said and done, the central issue is about graphics memory. The AGP vs. PCI issue gets dragged in because AGP is proposed as an alternative way to access more memory. Let me re-summarize my philosophy on this matter.
RULE #1 — Make sure that you have adequate graphics memory. If you have adequate graphics memory, you will not observe much of a performance difference between AGP and PCI. If you do not have adequate graphics memory, you are eventually going to run into texture related performance problems with PCI or with AGP. There are MANY variables that determine how big those problems are, and how they should best be solved.
Though I don’t think AGP is bad, relying on AGP’s execute mode texturing to recover from breaking Rule#1, is a less than ideal solution. It may potentially sacrifice MIPS, sacrifice accelerator performance, or both. If you already paid for AGP, and you inadvertently violated rule #1 in the process, AGP becomes a convenient, «no additional cost» way to deal with the problem. Convenient it is, but performance is another matter, and performance is not «free».
As I said, in order to be fully armed to make proper judgements, we will have to look at current and future requirements for texturing bandwidth and for the geometry stream. Let’s start with texturing.
