Calculate the TPS of the CMP 170HX 10 GB on local AI models

NVIDIA 10 GB HBM2e 1,560 GB/s September 2021

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

381 of 679 models it can run

Largest model it holds

Ling-lite-1.5 ("Bailing")

16.8B · Q3_K_M · 106 tok/s

Fastest model

Gemma 3 QAT 1B

661 tok/s · 1B

What AI models can a CMP 170HX 10 GB run?

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.

381 models match

Calculating
Quantisation Fit
661 tok/s

562–793

Gemma 3 1B 1B Mar 2025 1.8 GB 33k tokens Q8_0 Comfortable
661 tok/s

562–793

Gemma 3 QAT 1B 1B Apr 2025 1.8 GB 33k tokens Q8_0 Comfortable
661 tok/s

396–1,057 · low confidence

HGRN 1B (WT 103) 1B Nov 2023 1.8 GB 131k tokens ? Q8_0 Comfortable
661 tok/s

396–1,057 · low confidence

LLama 3..2 Typhoon 2 1B 1B Dec 2024 1.8 GB 131k tokens ? Q8_0 Comfortable
661 tok/s

396–1,057 · low confidence

OLMo-1B 1B Feb 2024 1.8 GB 131k tokens ? Q8_0 Comfortable
661 tok/s

396–1,057 · low confidence

Pythia-1b 1B Apr 2023 1.8 GB 131k tokens ? Q8_0 Comfortable
619 tok/s

371–990 · low confidence

DeepSeekMoE-16B 16B Jan 2024 8.1 GB 4k tokens Q3_K_M Tight
612 tok/s

367–979 · low confidence

OpenELM-1.1B 1.1B May 2024 1.9 GB 131k tokens ? Q8_0 Comfortable
601 tok/s

360–961 · low confidence

DeciCoder-1B 1.1B Aug 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
601 tok/s

360–961 · low confidence

SantaCoder 1.1B Jan 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
601 tok/s

360–961 · low confidence

TinyLlama-1.1B (1T token checkpoint) 1.1B Oct 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
601 tok/s

360–961 · low confidence

TinyLlama-1.1B (3T token checkpoint) 1.1B Oct 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
551 tok/s

330–881 · low confidence

EXAONE 4.0 (1.2B) 1.2B Jul 2025 2.0 GB 131k tokens ? Q8_0 Comfortable
551 tok/s

330–881 · low confidence

MinerU2.5 1.2B Sep 2025 2.0 GB 131k tokens ? Q8_0 Comfortable
551 tok/s

330–881 · low confidence

Pleias 1.0 1.2B 1.2B Dec 2024 2.0 GB 131k tokens ? Q8_0 Comfortable
551 tok/s

330–881 · low confidence

Pleias-RAG-1B 1.2B Apr 2025 2.0 GB 131k tokens ? Q8_0 Comfortable
537 tok/s

457–645

Llama 3.2 1B 1.2B Sep 2024 2.2 GB 131k tokens Q8_0 Comfortable
530 tok/s

318–847 · low confidence

MiniCPM-1.2B 1.2B Jun 2024 2.0 GB 131k tokens ? Q8_0 Comfortable
508 tok/s

305–813 · low confidence

DeepSeek Coder 1.3B 1.3B Jan 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
508 tok/s

305–813 · low confidence

DeepSeek-VL-1.3B 1.3B Mar 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
508 tok/s

305–813 · low confidence

DigiRL 1.3B Jun 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
508 tok/s

305–813 · low confidence

GLA Transformer 1.3B 1.3B Aug 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
508 tok/s

305–813 · low confidence

Janus 1.3B 1.3B Oct 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
508 tok/s

305–813 · low confidence

Kosmos-2.5 1.3B Aug 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
508 tok/s

305–813 · 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

CMP 170HX 10 GB 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
10 GB
Memory bandwidth
1,560 GB/s
Memory type
HBM2e
Memory bus width
5,120 bit
Memory clock
1.22 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
GA100
Architecture
Ampere
Generation
Mining GPUs
Foundry
TSMC
Process size
7 nm
Transistors
54.2 billion
Transistor density
65,600 K/mm²
Die size
826 mm²
Package
BGA-2743
Released
1 September 2021

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
1.14 GHz
Boost clock
1.41 GHz

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
4,480
Texture mapping units
280
Render output units
128
Streaming multiprocessors
70
Tensor cores
280
L1 cache
192 KB
L2 cache
10 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.

Half precision (FP16)
50.5 TFLOPS
Single precision (FP32)
12.6 TFLOPS
Double precision (FP64)
6.3 TFLOPS
Pixel rate
181 GPixel/s
Texture rate
395 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)
250 W
Suggested power supply
600 W
Power connectors
2x 8-pin
Bus interface
PCIe 1.0 x4
Slot width
IGP

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.

CUDA compute capability
8.0
OpenCL
3.0

Listings

Where to buy a CMP 170HX 10 GB

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

Memory: the specification that decides everything

Memory

10 GB

Bandwidth

1,560 GB/s

Largest model

Ling-lite-1.5 ("Bailing")

At 10 GB of HBM2e the CMP 170HX 10 GB is limited to the smaller end of the catalogue. About 9 GB is actually available to a runtime, and a model has to fit entirely inside it before generating anything at all.

Bandwidth is 1,560 GB/s across a 5,120-bit bus. Generating a token means reading every weight once, so that figure sets the pace more than any other number here, and at this level text arrives faster than most people read.

The figure is the memory clock — 1.22 GHz here — multiplied by the bus width. It is why core counts predict generation speed so poorly.

Put together, the largest model that fits is Ling-lite-1.5 ("Bailing") at 16.8B, running Q3_K_M and producing around 106 tokens per second.

The chip and how it was built

The CMP 170HX 10 GB is built on the GA100 graphics processor, using NVIDIA's Ampere architecture, as part of the Mining GPUs generation.

The chip is manufactured by TSMC, on a 7 nm process, with a die measuring 826 mm², holding 54.2 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 September 2021, roughly 4 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

FP16

50.5 TFLOPS

FP64

6.3 TFLOPS

Tensor cores

280

On paper the CMP 170HX 10 GB reaches 50.5 TFLOPS at half precision and 12.6 TFLOPS at single precision. These are peak figures no real workload sustains, and generating text reaches only a small fraction of them — decoding is limited by memory rather than arithmetic, which is why a card can look enormously powerful here and still produce tokens at an ordinary rate.

Double-precision throughput is 6.3 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.

The card carries 280 tensor cores across 70 streaming multiprocessors. These accelerate the matrix arithmetic at the heart of a transformer, and they are what make prompt processing — reading a long document before answering — dramatically faster than it would otherwise be.

Clocks run from 1.14 GHz at base to 1.41 GHz 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 CMP 170HX 10 GB has 192 KB of L1 cache, backed by 10 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 4,480 shading units, 280 texture mapping units, and 128 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

250 W

The CMP 170HX 10 GB is rated at 250 W, with a 600 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 igp, and needs 2x 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 1.0 x4. 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 a CMP 170HX 10 GB can run

The biggest open-weight models that fit on this card, newest first. Each is shown at the best compression the card can hold.

  1. 01 Ring-mini-linear-2.0 16.4B · Q3_K_M · Oct 2025 109 tok/s
  2. 02 Ling-mini-base-2.0-20T 16B · Q3_K_M · Sep 2025 111 tok/s
  3. 03 Ling-lite-1.5 ("Bailing") 16.8B · Q3_K_M · Mar 2025 106 tok/s
  4. 04 Nanbeige2-16B-Chat 15.8B · Q3_K_M · May 2024 113 tok/s
  5. 05 DeepSeekMoE-16B 16B · Q3_K_M · Jan 2024 619 tok/s
  6. 06 Nanbeige-16B 16B · Q3_K_M · Nov 2023 111 tok/s
  7. 07 CodeT5+ 16B · Q3_K_M · May 2023 111 tok/s
  8. 08 CodeGen2 16B · Q3_K_M · May 2023 111 tok/s
  9. 09 MOSS-Moon-003 16B · Q3_K_M · Apr 2023 111 tok/s
  10. 10 CodeGen-Mono 16.1B 16.1B · Q3_K_M · Feb 2023 111 tok/s

The fastest AI models on a CMP 170HX 10 GB

Where this card produces tokens quickest. Smaller models dominate here, because generating each token means reading the whole model out of memory once.

  1. 01 Gemma 3 QAT 1B 1B · Q8_0 · 1.8 GB 661 tok/s
  2. 02 Gemma 3 1B 1B · Q8_0 · 1.8 GB 661 tok/s
  3. 03 LLama 3..2 Typhoon 2 1B 1B · Q8_0 · 1.8 GB 661 tok/s
  4. 04 OLMo-1B 1B · Q8_0 · 1.8 GB 661 tok/s
  5. 05 HGRN 1B (WT 103) 1B · Q8_0 · 1.8 GB 661 tok/s
  6. 06 Pythia-1b 1B · Q8_0 · 1.8 GB 661 tok/s
  7. 07 DeepSeekMoE-16B 16B · Q3_K_M · 8.1 GB 619 tok/s
  8. 08 OpenELM-1.1B 1.1B · Q8_0 · 1.9 GB 612 tok/s
  9. 09 TinyLlama-1.1B (1T token checkpoint) 1.1B · Q8_0 · 1.9 GB 601 tok/s
  10. 10 TinyLlama-1.1B (3T token checkpoint) 1.1B · Q8_0 · 1.9 GB 601 tok/s

Step by step

How to work out the tokens per second of a CMP 170HX 10 GB

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.

  1. 01

    Start with the model, not the specification

    The table lists 381 models this CMP 170HX 10 GB can run. Search by name, or by size — typing 27b matches on the parameter count even when the name never states it.

  2. 02

    Set the context length you will actually use

    Longer conversations cost memory on top of the weights. With 10 GB to work in, that is frequently the difference between a model fitting and not.

  3. 03

    Pin the comparison to one quality level

    Compression is what lets bigger models fit. The quality control drops any model that needs more of it than you are willing to give.

  4. 04

    Look at the range, not just the number

    Each speed is an estimate for a single conversation, with a range beneath it — 661 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.

  5. 05

    Check the memory column before committing

    Compare what each model needs with the 10 GB this card provides. Tight means it works today; comfortable means it still works when the conversation grows.

  6. 06

    Check the same model from the other side

    Following a model through to its own page lists all the hardware that can run it, so you can see where the CMP 170HX 10 GB sits against the alternatives.

Answers

CMP 170HX 10 GB — common questions

01

How much power does a CMP 170HX 10 GB use?

The CMP 170HX 10 GB has a rated board power of 250 W, and a 600 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.

02

How much cache does a CMP 170HX 10 GB have?

The CMP 170HX 10 GB has 192 KB of L1 cache, and 10 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.

03

What are the TFLOPS of a CMP 170HX 10 GB?

The CMP 170HX 10 GB is rated at 50.5 TFLOPS at half precision and 12.6 TFLOPS at single precision. These are peak arithmetic ceilings rather than achievable rates, and text generation reaches only a small fraction of them because it is limited by memory bandwidth instead.

04

How many tensor cores does a CMP 170HX 10 GB have?

The CMP 170HX 10 GB has 280 tensor cores across 70 streaming multiprocessors. They accelerate the matrix arithmetic a transformer is built from, which mainly speeds up processing a long prompt rather than producing the reply.

05

Does the CMP 170HX 10 GB support CUDA?

Yes. The CMP 170HX 10 GB reports CUDA compute capability 8.0. Capability 7.0 and above has tensor cores, which modern inference software uses; below that it falls back to slower code paths for quantised models.

06

What bus interface does the CMP 170HX 10 GB use?

It uses PCIe 1.0 x4. 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.

07

Is the CMP 170HX 10 GB good for running local AI models?

Its memory limits it to smaller models and its bandwidth is high enough to generate text faster than most people read. In total it runs 381 of the models we track. Whether that is enough depends entirely on which model you want — the table above answers that directly.

08

Can a CMP 170HX 10 GB run a model that does not fit in its memory?

Offloading past the card's 10 GB is possible and usually a false economy: the system-memory portion is slow enough to dominate the result.

09

Would two CMP 170HX 10 GB cards be twice as fast?

Capacity adds, throughput does not. Two of them give you 20 GB to work with rather than twice the tokens per second — every figure here is for a single CMP 170HX 10 GB.

10

What AI models can a CMP 170HX 10 GB run?

381 of the 679 open-weight language models we track fit on a CMP 170HX 10 GB 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.

11

What is the largest AI model a CMP 170HX 10 GB can run?

The largest model in our catalogue that fits on a CMP 170HX 10 GB is Ling-lite-1.5 ("Bailing") at 16.8B parameters, compressed to Q3_K_M. It generates roughly 106 tokens per second and needs about 8.9 GB of the card's memory.

12

How many tokens per second does a CMP 170HX 10 GB produce?

It depends on the model. On a CMP 170HX 10 GB the fastest model we track is Gemma 3 QAT 1B at about 661 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.

13

Can a CMP 170HX 10 GB run a 7B model?

Yes. For example a CMP 170HX 10 GB runs DeepSeek Coder 6.7B at Q4_K_M, using about 8.3 GB of memory and generating around 228 tokens per second.

14

Can a CMP 170HX 10 GB run a 13B model?

Yes. For example a CMP 170HX 10 GB runs DeepSeekMoE-16B at Q3_K_M, using about 8.1 GB of memory and generating around 619 tokens per second.

15

How much memory does a CMP 170HX 10 GB have?

A CMP 170HX 10 GB has 10 GB of HBM2e memory. Around a tenth of that is reserved by the inference runtime and the driver, leaving roughly 9 GB available for a model and its conversation.

16

What is the memory bandwidth of a CMP 170HX 10 GB?

The CMP 170HX 10 GB has 1,560 GB/s of memory bandwidth, across a 5,120-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.

17

What type of memory does a CMP 170HX 10 GB use?

It uses HBM2e clocked at 1.22 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.

18

Who makes the CMP 170HX 10 GB?

The CMP 170HX 10 GB is a NVIDIA product, with the chip manufactured by TSMC, on a 7 nm process.

19

When was the CMP 170HX 10 GB released?

The CMP 170HX 10 GB was released in September 2021.

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.

All GPUs