Sign in with
Sign up | Sign in

Tom2D: Polygons

Part 2: 2D, Acceleration, And Windows: Aren't All Graphics Cards Equal?
By , Igor Wallossek

Next up, drawing polygons. Here, too, various test candidates demonstrate interesting weaknesses.

Polygon benchmarkPolygon benchmark

Test Content:

For this test, we draw a total of 60,000 polygons with anywhere from 3 to 30 vertices, various line weights, and either color-filled or empty in some randomly chosen position on screen. As with our other tests, the random number generator seed remains constant so all test runs produce the same results. All objects appear entirely within the display area, so no clipping is necessary.

GDI Functions Called

Direct drawing test for polygons:

CreatePen
CreateSolidBrush
GetStockObject
SelectObject
DeleteObject
Polygon

Polygons drawn from the DIB buffer:

CreateCompatibleDC
CreateDIBSection
Rectangle
BitBlt

Summary

When polygons get buffered, performance is the same whether Aero is turned on or off. For direct draw operations, the high-end graphics cards fall way behind the pack. We have to wonder what might be bottlenecking ATI's cards. Similarly, why is Nvidia's GeForce GTX 285 suffering from freezes and delays when rendering polygons? And why is it falling behind Intel's GMA 950/Atom combination and the older G45-based integrated core logic?

React To This Article