Calculate the TPS of the FirePro S9100 on local AI models
Every model in our catalogue assessed against this card at the context length and minimum quality you choose. Speed is an estimate for a single request, calculated from this card's memory bandwidth and the size of each model once compressed.
Calculated for this card
679 models in our catalogue altogether
Largest model it holds
ERNIE-4.5-21B-A3B
21B · Q3_K_M · 75.5 tok/s
Fastest model
Gemma 3 QAT 1B
106 tok/s · 1B
Which AI models can run on a FirePro S9100?
Set the inputs, read the answer
More context means more memory for the conversation cache. Speed is for a fresh conversation and does not change with this setting.
Hides models that would only fit by being compressed below this point.
396 models match
Calculating| Quantisation | Fit | ||||||
|---|---|---|---|---|---|---|---|
|
106
tok/s
63–169 · low confidence |
Gemma 3 1B | 1B | Mar 2025 | 1.8 GB | 33k tokens | Q8_0 | Comfortable |
|
106
tok/s
63–169 · low confidence |
Gemma 3 QAT 1B | 1B | Apr 2025 | 1.8 GB | 33k tokens | Q8_0 | Comfortable |
|
106
tok/s
63–169 · low confidence |
HGRN 1B (WT 103) ≈ | 1B | Nov 2023 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
106
tok/s
63–169 · low confidence |
LLama 3..2 Typhoon 2 1B ≈ | 1B | Dec 2024 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
106
tok/s
63–169 · low confidence |
OLMo-1B ≈ | 1B | Feb 2024 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
106
tok/s
63–169 · low confidence |
Pythia-1b ≈ | 1B | Apr 2023 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
97.9
tok/s
59–157 · low confidence |
OpenELM-1.1B ≈ | 1.1B | May 2024 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
96.1
tok/s
58–154 · low confidence |
DeciCoder-1B ≈ | 1.1B | Aug 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
96.1
tok/s
58–154 · low confidence |
SantaCoder ≈ | 1.1B | Jan 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
96.1
tok/s
58–154 · low confidence |
TinyLlama-1.1B (1T token checkpoint) ≈ | 1.1B | Oct 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
96.1
tok/s
58–154 · low confidence |
TinyLlama-1.1B (3T token checkpoint) ≈ | 1.1B | Oct 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
88.1
tok/s
53–141 · low confidence |
EXAONE 4.0 (1.2B) ≈ | 1.2B | Jul 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
88.1
tok/s
53–141 · low confidence |
MinerU2.5 ≈ | 1.2B | Sep 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
88.1
tok/s
53–141 · low confidence |
Pleias 1.0 1.2B ≈ | 1.2B | Dec 2024 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
88.1
tok/s
53–141 · low confidence |
Pleias-RAG-1B ≈ | 1.2B | Apr 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
86.0
tok/s
52–138 · low confidence |
Llama 3.2 1B | 1.2B | Sep 2024 | 2.2 GB | 131k tokens | Q8_0 | Comfortable |
|
84.7
tok/s
51–136 · low confidence |
DeepSeekMoE-16B | 16B | Jan 2024 | 10.0 GB | 4k tokens | Q4_K_M | Tight |
|
84.7
tok/s
51–136 · low confidence |
MiniCPM-1.2B ≈ | 1.2B | Jun 2024 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
81.3
tok/s
49–130 · low confidence |
DeepSeek Coder 1.3B ≈ | 1.3B | Jan 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
81.3
tok/s
49–130 · low confidence |
DeepSeek-VL-1.3B ≈ | 1.3B | Mar 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
81.3
tok/s
49–130 · low confidence |
DigiRL ≈ | 1.3B | Jun 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
81.3
tok/s
49–130 · low confidence |
GLA Transformer 1.3B ≈ | 1.3B | Aug 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
81.3
tok/s
49–130 · low confidence |
Janus 1.3B ≈ | 1.3B | Oct 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
81.3
tok/s
49–130 · low confidence |
Kosmos-2.5 ≈ | 1.3B | Aug 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
81.3
tok/s
49–130 · low confidence |
Otter ≈ | 1.3B | May 2023 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
Speeds are estimates for a single request — one conversation at a time — calculated from memory bandwidth, model size and quantisation. Real throughput varies with the inference runtime and its version. Figures published by hardware vendors measure many simultaneous requests and are much higher.
On record
FirePro S9100 full specification
Everything on record for this board, ordered by how much it bears on running a language model rather than by how a spec sheet would list it. Memory comes first because it decides the outcome; the rest is context.
Memory
The two specifications that decide what this card can run and how quickly. Capacity sets which models fit; bandwidth sets how many tokens per second they produce once they do.
- Memory size
- 12 GB
- Memory bandwidth
- 320 GB/s
- Memory type
- GDDR5
- Memory bus width
- 512 bit
- Memory clock
- 1.25 GHz
The chip
Which processor is on the board and how it was manufactured. A smaller process size generally means more performance for the same power.
- Graphics processor
- Hawaii
- Architecture
- GCN 2.0
- Generation
- FirePro Server(Sx100)
- Foundry
- TSMC
- Process size
- 28 nm
- Transistors
- 6.2 billion
- Transistor density
- 14,200 K/mm²
- Die size
- 438 mm²
- Released
- 2 October 2014
Clock speeds
How fast the processor runs. Worth far less here than on a gaming benchmark: generating text is limited by memory bandwidth, so a higher clock barely moves the result.
- Base clock
- 824 MHz
- Boost clock
- 824 MHz
Processing units
What the chip contains. These drive graphics performance and matter mainly for processing a long prompt rather than for producing the answer.
- Shading units
- 2,560
- Texture mapping units
- 160
- Render output units
- 64
- L1 cache
- 16 KB
- L2 cache
- 1 MB
Theoretical performance
Peak arithmetic rates published for the board. These are ceilings that no real workload reaches, and generating text reaches a small fraction of them because it is limited by memory rather than arithmetic.
- Single precision (FP32)
- 4.2 TFLOPS
- Double precision (FP64)
- 2.1 TFLOPS
- Pixel rate
- 53 GPixel/s
- Texture rate
- 132 GTexel/s
The board
What it takes to physically install and power the card — the practical constraints that decide whether it fits the machine you already own.
- Power draw (TDP)
- 225 W
- Suggested power supply
- 550 W
- Power connectors
- 1x 6-pin + 1x 8-pin
- Bus interface
- PCIe 3.0 x16
- Slot width
- Dual-slot
- Dimensions
- 267 mm
Software support
Which graphics and compute interfaces the card supports. CUDA compute capability is the one that bears on inference: below 7.0 there are no tensor cores, and modern inference software falls back to slower code paths.
- DirectX
- 12.0
- OpenGL
- 4.6
- Vulkan
- 1.2
- OpenCL
- 2.1
- Shader model
- 6.5
Listings
Where to buy a FirePro S9100
No vendor is currently listing this card. Listings come from vendors who publish them here directly — browse the vendor directory to see who is selling what.
What the numbers mean
Capacity and bandwidth
Memory
12 GB
Bandwidth
320 GB/s
Largest model
ERNIE-4.5-21B-A3B
12 GB of GDDR5 puts the FirePro S9100 comfortably into small and mid-sized models, with roughly 10.8 GB usable once the driver overhead is taken out. The largest models are out of reach without splitting them.
The memory bus moves 320 GB/s across a 512-bit bus. That is the number that governs generation speed — arithmetic per byte read is small enough that the bus, not the cores, is what everything waits on.
That comes from a 1.25 GHz memory clock across the bus width above. Widening the bus and raising the clock are the two levers a manufacturer has, which is why a card with unremarkable cores can still generate quickly.
Put together, the largest model that fits is ERNIE-4.5-21B-A3B at 21B, running Q3_K_M and producing around 75.5 tokens per second.
The chip and how it was built
The FirePro S9100 is built on the Hawaii graphics processor, using AMD's GCN 2.0 architecture, as part of the FirePro Server(Sx100) generation.
The chip is manufactured by TSMC, on a 28 nm process, with a die measuring 438 mm², holding 6.2 billion transistors. A smaller process generally means more performance for the same power, though for language models it matters far less than the memory subsystem.
It was released in October 2014, roughly 11 years ago. Inference software support tends to follow hardware by a year or two, so a card of this age generally has mature, well-optimised code paths available to it.
Compute throughput, and why it matters less than it looks
FP64
2.1 TFLOPS
Double-precision throughput is 2.1 TFLOPS. It has no bearing on running a language model — no inference runtime uses it — but it separates datacentre parts from consumer ones, since the latter deliberately restrict it.
Clocks run from 824 MHz at base to 824 MHz boosted. Worth far less here than on a gaming benchmark: raising the clock speeds up the arithmetic, and the arithmetic is not what generation is waiting on.
Cache and processing units
The FirePro S9100 has 16 KB of L1 cache, backed by 1 MB of L2. Cache absorbs a share of the memory traffic that would otherwise hit the main bus, which is the one place on this page where a number other than bandwidth quietly affects generation speed — a large L2 lets more of the working set stay close to the cores.
There are 2,560 shading units, 160 texture mapping units, and 64 render output units. These drive graphics workloads and contribute to prompt processing, but they sit idle for much of the time a model spends generating a reply.
Power, size and installation
Power draw
225 W
The FirePro S9100 is rated at 225 W, with a 550 W power supply suggested for the whole system. Running a language model keeps a card busy in bursts rather than continuously — it draws hard while generating and idles between requests — so sustained draw over a working day is usually well below the rated figure.
The board occupies a dual-slot, measuring 267 mm long, and needs 1x 6-pin + 1x 8-pin. Worth checking against the case and power supply already in the machine, since the largest cards need considerably more of both than a typical desktop provides.
It connects over PCIe 3.0 x16. The interface governs how quickly a model is loaded from disk into the card, not how fast it runs once there, so a narrower link costs a few seconds at startup and nothing thereafter.
The extremes
The largest AI models that run on a FirePro S9100
The biggest open-weight models that fit on this card, newest first. Each is shown at the best compression the card can hold.
The fastest AI models on a FirePro S9100
Where this card produces tokens quickest. Smaller models dominate here, because generating each token means reading the whole model out of memory once.
Step by step
How to work out the tokens per second of a FirePro S9100
You do not have to calculate anything by hand — the gputps.com calculator on this page has already worked it out for every model this card can hold. Reading off the answer takes six steps.
-
01
Find the model in the table
The table lists 396 models this FirePro S9100 can run. Search by name, or by size — typing 27b matches on the parameter count even when the name never states it.
-
02
Match the context to your work
Set the context to your real working length. Short questions cost almost nothing; a long document can consume a large share of the card's 12 GB.
-
03
Choose how far you will compress
Compression is what lets bigger models fit. The quality control drops any model that needs more of it than you are willing to give.
-
04
Read the speed and the range
Each speed is an estimate for a single conversation, with a range beneath it — 106 tok/s on Gemma 3 QAT 1B at the top end here. The same card and model vary by thirty to fifty per cent between inference runtimes.
-
05
Check the memory column before committing
A tight fit runs but leaves no room to raise the context later; comfortable has headroom. The memory column shows what each model needs against the 12 GB available.
-
06
Open the model to compare cards
Following a model through to its own page lists all the hardware that can run it, so you can see where the FirePro S9100 sits against the alternatives.
Answers
FirePro S9100 — common questions
Does the FirePro S9100 support CUDA?
No. CUDA is NVIDIA-only, and the FirePro S9100 is a AMD card. It runs language models through ROCm, Vulkan or Metal depending on the software, which are less mature than the CUDA path — our estimates apply a penalty for that.
What bus interface does the FirePro S9100 use?
It uses PCIe 3.0 x16. This governs how fast a model is loaded onto the card rather than how fast it runs once loaded, so it costs a few seconds at startup and nothing during generation.
Is the FirePro S9100 good for running local AI models?
Its memory covers small and mid-sized models, though the largest are out of reach though its bandwidth means generation will feel slow on larger models. In total it runs 396 of the models we track. Whether that is enough depends entirely on which model you want — the table above answers that directly.
Can a FirePro S9100 run a model that does not fit in its memory?
Only partly. Layers beyond the 12 GB sit in system memory and run at a fraction of the speed, so a mostly-offloaded model is rarely worth using. Every figure here assumes it is fully resident on the card.
Would two FirePro S9100 cards be twice as fast?
Capacity adds, throughput does not. Two of them give you 24 GB to work with rather than twice the tokens per second — every figure here is for a single FirePro S9100.
What AI models can a FirePro S9100 run?
396 of the 679 open-weight language models we track fit on a FirePro S9100 and can be run locally on it. The table on this page lists every one, with the memory it needs, the quantisation it runs at and an estimated generation speed.
What is the largest AI model a FirePro S9100 can run?
The largest model in our catalogue that fits on a FirePro S9100 is ERNIE-4.5-21B-A3B at 21B parameters, compressed to Q3_K_M. It generates roughly 75.5 tokens per second and needs about 10.1 GB of the card's memory.
How many tokens per second does a FirePro S9100 produce?
It depends on the model. On a FirePro S9100 the fastest model we track is Gemma 3 QAT 1B at about 106 tokens per second, while larger models run proportionally slower because each token requires reading the whole model out of memory once. Speeds are estimates for a single conversation at a time.
Can a FirePro S9100 run a 7B model?
Yes. For example a FirePro S9100 runs DeepSeek Coder 6.7B at Q6_K, using about 9.9 GB of memory and generating around 22.9 tokens per second.
Can a FirePro S9100 run a 13B model?
Yes. For example a FirePro S9100 runs DeepSeekMoE-16B at Q4_K_M, using about 10.0 GB of memory and generating around 84.7 tokens per second.
How much memory does a FirePro S9100 have?
A FirePro S9100 has 12 GB of GDDR5 memory. Around a tenth of that is reserved by the inference runtime and the driver, leaving roughly 10.8 GB available for a model and its conversation.
What is the memory bandwidth of a FirePro S9100?
The FirePro S9100 has 320 GB/s of memory bandwidth, across a 512-bit memory bus. This is the single best predictor of how fast it generates text, because producing each token means reading the entire model out of memory once.
What type of memory does a FirePro S9100 use?
It uses GDDR5 clocked at 1.25 GHz. HBM types are found on datacentre accelerators and carry far more bandwidth than the GDDR used on desktop cards, which is why they generate tokens considerably faster at the same capacity.
Who makes the FirePro S9100?
The FirePro S9100 is a AMD product, with the chip manufactured by TSMC, on a 28 nm process.
When was the FirePro S9100 released?
The FirePro S9100 was released in October 2014.
How much power does a FirePro S9100 use?
The FirePro S9100 has a rated board power of 225 W, and a 550 W system power supply is suggested. Generating text draws hard in bursts and idles between requests, so average consumption over a working session is normally well below the rated figure.
How much cache does a FirePro S9100 have?
The FirePro S9100 has 16 KB of L1 cache, and 1 MB of L2 cache. Cache absorbs part of the memory traffic that would otherwise reach the main bus, so a larger L2 gives a modest lift to generation speed beyond what bandwidth alone predicts.
The other direction
Looking at it from the other side?
This page starts from the hardware. If you already know which model you want and need to know what it takes to run it, start from the model instead.