System setup and overclocking
|
Software
All benchmarks were conducted with vertical sync' disabled. Benchmarks were run 3 times consecutively, an average score was taken. The host system was configured for maximum performance. A fresh installation of Windows XP was used in both instances. I'm taking a little different approach in the benchmarks that follow. I have been benchmarking with a 2.4GHz processor, but it stands to reason that prospective buyers of Ti 4200s will have differing host systems. What I've done, therefore, is to conduct the same benchmarks at 1600MHz (let's call this low-end) and 2533MHz (premium), just to illustrate how heavily the CPU influences the overall results. I've simply overclocked my 1.6A to 2.533GHz (158FSB). It does this at default voltage as it's an excellent overclocker. A quick inspection with our preferred overclocking utility, Rivatuner, reveals that the ABIT Siluro Ti 4200 plays it firmly by the book.
Unlike a couple of other cards, the operating frequencies are bang-on specification. Overclocking Unless a manufacturer makes an overt effort to go over and above the reference design's cooling, you're overclocking is more down to luck than anything else. You could receive identical 10 cards from the same manufacturer and each would overclock differently, that's why I don't pay too much attention to the results obtained from a sample of one. To test the ability of the core, I ran the Dragothic benchmark from 3DMark 2001SE, it's heavy use of shaders puts an onus on the GPU. Going up in 5MHz increments, and running the benchmark 3 times consecutively along the way, I got to a stable core speed of 285MHz, or 35MHz above Ti 4200 specification and 10MHz above Ti 4400 specification. Whilst this is no match for the MSI Ti 4200's 320MHz core, it again shows that luck is a factor. The very quiet cooler on the Siluro Ti 4200 keeps the noise down but perhaps can't cut it at high core speeds. I noted earlier that the memory on the Siluro Ti 4200 was surprisingly on specification as it used 4ns RAM (500MHz DDR). I was quite pleased when I arrived at 590MHz memory with no artifacts after an extended run in 3DMark. I must note that the memory did become extremely warm to the touch, perhaps RAMsinks would have helped a little here. 285MHz core and 590MHz memory speeds put this Ti 4200 above the default specifications of the more expensive Ti 4400. I must temper this by stating that the latter is equipped with 128MB of on-board memory so you cannot always directly compare the two. Anyway, on to the benchmarks. I have no comparative benchmarks from other cards as this is the first time that I'm trying a new method of benchmarking (using both 1600MHz and 2533MHz). The reason that I'm going for these speeds is that they represent the two ends of the spectrum. I see a number of reviews done exclusively with 2.53GHz P4s and XP2100s. The majority of us don't have these processors and therefore are more interested in the performance of the cards with slightly more common processors. Hence using both 1600MHz and 2533MHz (1.6A overclocked). I'm also now using the 29.42 Detonator drivers as opposed to the 28.32 for the majority of other cards tested. Including their results would not be scientifically acceptable. |
