ATI Radeon X850 XT AGP

The ATI Radeon X850 XT AGP is an AMD-manufactured graphics card built around the R481 chip and the R400 architecture, placed in the Radeon R400 AGP (X850) generation. The record does not assign a series name or codename to this product. It was released on 2005-02-27 and is marked end-of-life. TSMC fabricated the chip on a 130 nm process, packing 160 million transistors into a 297 mm² die, for a transistor density of 538.7K per mm². The card connects through AGP 8x, uses a dual-slot bracket, requires one Molex power connector, and carries a 69 W TDP with a 250 W suggested PSU. Its memory subsystem is 256 MB of GDDR3 on a 256-bit bus, rated at 34.56 GB/s. The benchmark array for this product is empty, the nearestRivals list is empty, and the only aggregate position stored is a percentileVsAllGpus value of 50.

Benchmark Performance

The benchmarks array for this product contains no entries. The avgBenchmarkScore field is recorded as 0, which in this dataset is the value that appears when no per-workload scores are available rather than a measured average. The sole distribution metric is percentileVsAllGpus, and it is 50. That is the 50th percentile in the database's all-GPU ranking, but the percentile is not attached to any individual score in the record.

No nearestRivals entries accompany the card, so no deltaPct values can be reported. Without rival scores, exact percentage deltas — such as one card being a certain percent ahead of or behind another — cannot be computed from the FACT PACK. This is a data limitation, not a performance verdict.

What the record does contain is the fixed-function pipeline specification. The chip has 16 texture mapping units and 16 render output units, producing a texture rate of 8.320 GTexel/s and a pixel rate of 8.320 GPixel/s. Those are the maximum listed rates for texture and pixel processing. The memory subsystem supplies 34.56 GB/s of bandwidth to feed those units. No shading-unit count is listed, and the FP32 and FP16 throughput fields are null, so general-purpose compute throughput cannot be quantified from this record. The base, boost, and game clock fields are also null; the only clock listed is the memory clock of 540 MHz, with a 1080 Mbps effective transfer rate.

Ray Tracing and Feature Set

The rtCores and tensorCores fields are both null. The record therefore provides no dedicated ray tracing core count and no tensor core count for this card. In terms of API support, the card lists DirectX 9.0b (9_2) and OpenGL 2.1. No Vulkan version is listed. The absence of Vulkan is recorded as null in the API table, so Vulkan-based workloads are not covered by this entry.

The feature set otherwise consists of the display and connectivity options: 1x DVI, 1x VGA, and 1x S-Video output. The physical feature list includes a dual-slot form factor, a single Molex power connector, and an AGP 8x bus interface. The TDP is 69 W, with a suggested PSU of 250 W. The R400 architecture and R481 chip are the defining architectural features, with the 130 nm TSMC process and the 297 mm² die as the implementation details.

How It Compares

The nearestRivals array is empty. There are no rival names, no rival scores, and no deltaPct values in the FACT PACK, so a rival-by-rival comparison cannot be constructed from this data. This section would normally state, for each listed rival, the percentage delta in benchmark scores; in this record, no such entries exist.

The only comparative anchors are the generation fields. The predecessor is the Radeon R300. The successor is the Radeon R500 AGP. The card itself belongs to the Radeon R400 AGP (X850) generation. These fields establish position between the R300 and R500 AGP product families, but they do not provide numeric performance relationships.

The aggregate percentile field is the only quantitative ranking signal: percentileVsAllGpus is 50. That places the card in the middle of the all-GPU ranking in the database. Because the benchmarks array is empty and avgBenchmarkScore is 0, the percentile is a standalone positional value rather than a computed result from a set of scores.

FAQ

Q: What chip and architecture does the ATI Radeon X850 XT AGP use?

A: It uses the R481 chip on the R400 architecture. TSMC builds the chip on a 130 nm process, with 160 million transistors on a 297 mm² die and a transistor density of 538.7K per mm².

Q: How much memory does the card have, and what kind?

A: It has 256 MB of GDDR3 memory on a 256-bit bus. The memory clock is 540 MHz, with an effective 1080 Mbps data rate, yielding 34.56 GB/s of bandwidth.

Q: What graphics APIs are supported?

A: The card supports DirectX 9.0b (9_2) and OpenGL 2.1. No Vulkan API version is listed. The rtCores and tensorCores fields are null, so no dedicated ray tracing or tensor core counts are recorded.

Q: What are the card's fill rates?

A: With 16 TMUs and 16 ROPs, the card reaches 8.320 GTexel/s texture rate and 8.320 GPixel/s pixel rate.

Q: What power and physical requirements are listed?

A: The TDP is 69 W, and the suggested PSU is 250 W. Power is supplied through a single Molex connector, and the card uses a dual-slot bracket. It connects through an AGP 8x interface.

Q: When was it released, and what is its production status?

A: It was released on 2005-02-27 and is marked end-of-life. Its predecessor is the Radeon R300, and its successor is the Radeon R500 AGP.

Memory Subsystem

The memory subsystem is built around a 256 MB GDDR3 frame buffer connected over a 256-bit memory bus. The listed bandwidth is 34.56 GB/s, with a memory clock of 540 MHz and an effective 1080 Mbps data rate. These figures — capacity, bus width, and bandwidth — define the memory envelope for all workloads.

At high resolutions, the relevant constraint is the combination of capacity and bandwidth. A 256 MB buffer holds a limited amount of frame, texture, and geometry data; beyond that, data must be re-fetched or compressed. The 256-bit bus and 34.56 GB/s bandwidth set the speed at which that data can move. No benchmark scores exist in the record to quantify how this memory configuration behaves at specific resolutions, but the memory parameters themselves are explicit.

The fill-rate side ties into the same memory path. With 16 ROPs and a pixel output rate of 8.320 GPixel/s, the memory subsystem must keep those pixel writes supplied through the 256-bit interface. The texture path, rated at 8.320 GTexel/s, also draws on the same 34.56 GB/s bandwidth. The memory clock's 1080 Mbps effective rate is the data rate field attached to the 256-bit bus. No benchmark data is present to show how this memory configuration performs at high resolutions; the capacity and bandwidth are the only quantitative constraints available in the record.