DLDL (PASCAL) TPS calculator

Open weights University of Oxford 564M parameters November 2016

Each card below is assessed against this model at the context length and minimum quality you choose. Speed is an estimate for a single request, calculated from the card's memory bandwidth and the size of the model once compressed.

Calculated for this model

818 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Tesla C1080

4 GB · Q8_0 · 65.4 tok/s

Fastest card

B200

6,008 tok/s · 180 GB

Which GPUs can run DLDL (PASCAL)?

Set the inputs, read the answer

Context length tokens

A longer conversation needs more memory, which can push this model off smaller cards.

Minimum quality

Hides cards that would only fit the model by compressing it below this point.

Search cards

Card memory

Also show cards that cannot run it

818 cards match

Calculating

					Needs	Quantisation	Fit
6,008 tok/s 3,605–9,612 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	1.3 GB	Q8_0	Comfortable
6,008 tok/s 3,605–9,612 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	1.3 GB	Q8_0	Comfortable
4,797 tok/s 2,878–7,675 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	1.3 GB	Q8_0	Comfortable
4,797 tok/s 2,878–7,675 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	1.3 GB	Q8_0	Comfortable
3,837 tok/s 2,302–6,138 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	1.3 GB	Q8_0	Comfortable
3,672 tok/s 2,203–5,875 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.3 GB	Q8_0	Comfortable
3,672 tok/s 2,203–5,875 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.3 GB	Q8_0	Comfortable
3,514 tok/s 2,109–5,623 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	1.3 GB	Q8_0	Comfortable
3,119 tok/s 1,871–4,990 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	1.3 GB	Q8_0	Comfortable
3,119 tok/s 1,871–4,990 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	1.3 GB	Q8_0	Comfortable
3,119 tok/s 1,871–4,990 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	1.3 GB	Q8_0	Comfortable
2,959 tok/s 1,775–4,734 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	1.3 GB	Q8_0	Comfortable
2,523 tok/s 1,514–4,037 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.3 GB	Q8_0	Comfortable
2,523 tok/s 1,514–4,037 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	1.3 GB	Q8_0	Comfortable
2,523 tok/s 1,514–4,037 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	1.3 GB	Q8_0	Comfortable
2,523 tok/s 1,514–4,037 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.3 GB	Q8_0	Comfortable
2,523 tok/s 1,514–4,037 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	1.3 GB	Q8_0	Comfortable
1,921 tok/s 1,153–3,074 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	1.3 GB	Q8_0	Comfortable
1,921 tok/s 1,153–3,074 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	1.3 GB	Q8_0	Comfortable
1,601 tok/s 961–2,562 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	1.3 GB	Q8_0	Comfortable
1,567 tok/s 940–2,507 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	1.3 GB	Q8_0	Comfortable
1,532 tok/s 919–2,451 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	1.3 GB	Q8_0	Comfortable
1,532 tok/s 919–2,451 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	1.3 GB	Q8_0	Comfortable
1,532 tok/s 919–2,451 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	1.3 GB	Q8_0	Comfortable
1,532 tok/s 919–2,451 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	1.3 GB	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

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation: University of Oxford
Organisation type: Academia
Country: United Kingdom of Great Britain and Northern Ireland
Published: 6 November 2016
Authors: Bin-Bin Gao, Chao Xing, Chen-Wei Xie, Jianxin Wu, Xin Geng

What it does

The problem areas the model was built for. A model can carry several of each.

Domain: Vision
Task: Image classification
Approach: Supervised
Base model: VGG16,VGG19

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Parameters: 564M
Training data: 22,531 tokens

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Training hardware: NVIDIA Tesla K40c
Chips used: 1
Power draw: 283 W

Availability

Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.

Weights: Open — downloadable
Model access: Open weights (unrestricted)

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Why it is tracked: SOTA improvement
Record confidence: Likely

Sources

Where this record came from and when it was last checked.

Reference: Deep Label Distribution Learning With Label Ambiguity
Last updated: 11 February 2026

The extremes

The ten fastest GPUs that run DLDL (PASCAL)

Ranked by estimated tokens per second, newest card first where speeds tie. Because generation is bound by memory bandwidth, this ordering follows bandwidth rather than any gaming benchmark.

The smallest GPUs that still run DLDL (PASCAL)

The cheapest route in, by memory capacity. A tight fit runs the model but leaves nothing spare for a longer conversation.

What the numbers mean

What it takes to run this model

Minimum card

Tesla C1080

Memory needed

1.3 GB

Fastest

6,008 tok/s

DLDL (PASCAL) is small enough at 564M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The smallest card that holds it is the Tesla C1080 with 4 GB, running it at Q8_0 and producing around 65.4 tokens per second.

The quickest result comes from a B200 at around 6,008 tokens per second — its 8,000 GB/s of bandwidth is what buys that.

Background

DLDL (PASCAL) was published by University of Oxford, in United Kingdom of Great Britain and Northern Ireland, in November 2016. The organisation is categorised as academia.

It works in Vision, and is recorded as doing image classification.

It is derived from VGG16,VGG19 rather than trained from scratch, which is the usual way a specialised model is produced.

The weights being open is what puts this page in the calculator rather than only in the catalogue: it is a model you can actually hold.

Reading the throughput figures

Half the cards that hold it manage more than 168.7 tokens per second, and 809 exceed reading speed outright.

It is a dense model, so every parameter is read for every token produced. That makes speed track memory bandwidth almost exactly — a card with twice the bandwidth generates roughly twice as fast.

Its internal architecture is not on file, so memory is approximated from the parameter count and marked accordingly. Expect the real figure to differ, more so at long context.

Training and provenance

It was trained on about 22,531 tokens of text.

Its inclusion criterion is sOTA improvement.

Step by step

How to choose a GPU for DLDL (PASCAL)

The table above has already assessed every card we hold specifications for against this model. Getting to your answer takes six steps.

01

Read the memory figure first

Every card here has been checked against DLDL (PASCAL) — around 1.3 GB at Q8_0. Capacity is the gate — a card either holds it or it does not.
02

Match the context to your actual use

Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason DLDL (PASCAL) stops fitting a card that seemed fine.
03

Set a quality floor

Each card runs the least-compressed copy it can hold — Q8_0 on the smallest card that fits. Setting a floor drops the cards that only manage DLDL (PASCAL) by squeezing it further than you would want.
04

Rank by throughput rather than spec sheet

Sort by speed to see how cards rank for DLDL (PASCAL). It will not match a gaming ordering — generation is bound by memory bandwidth, which is why the B200 tops it at 6,008 tok/s.
05

Read the fit column last

Tight means DLDL (PASCAL) loads and works, with no room to raise the context later. Comfortable means you can. The difference matters more than a few tokens per second.
06

See what else that card runs

Following a card through to its own page shows every other model it can hold, which is the question that follows once DLDL (PASCAL) is settled.

Answers

DLDL (PASCAL) — common questions

Can I run DLDL (PASCAL) if it does not fit in my GPU?

Only by offloading, which is usually a false economy: the part in system memory drags the whole thing down. Our figures for DLDL (PASCAL) assume it is fully resident.

Would two GPUs run DLDL (PASCAL) faster?

A second card roughly doubles the memory available but not the generation rate. With 818 cards already able to run DLDL (PASCAL) alone, the case for pairing is weak.

Why does the quantisation differ between cards for DLDL (PASCAL)?

Each card is shown running the least-compressed copy it can hold, and DLDL (PASCAL) appears at 1 different compression levels across the cards that fit it. Bigger cards get the more accurate version.

How accurate are these DLDL (PASCAL) speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 3,605–9,612 tok/s on the B200, for instance. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

What GPU do I need to run DLDL (PASCAL)?

The smallest card in our catalogue that holds DLDL (PASCAL) is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.3 GB, and produces roughly 65.4 tokens per second. 818 cards in total can run it.

How fast is DLDL (PASCAL) on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 6,008 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 809 of the cards that can run DLDL (PASCAL) clear that.

How much VRAM does DLDL (PASCAL) need?

About 1.3 GB at Q8_0 compression, which is what the smallest card that runs it uses. Less compression needs more: the figures in the memory column above are recalculated for each card, because each one holds the least-compressed version it can.

Can I run DLDL (PASCAL) on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 1.3 GB and generating roughly 1,119 tokens per second — a comfortable fit.

Can I run DLDL (PASCAL) on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 1.3 GB and generating roughly 685 tokens per second — a comfortable fit.

Can I run DLDL (PASCAL) on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 1.3 GB and generating roughly 849 tokens per second — a comfortable fit.

Can I run DLDL (PASCAL) on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 1.3 GB and generating roughly 1,006 tokens per second — a comfortable fit.

Is DLDL (PASCAL) open source?

Its weights are published, so DLDL (PASCAL) can be downloaded and run on your own hardware. Note that open weights is not the same as open source in the full sense — it says nothing about the training data, the training code, or the commercial terms attached.

How many parameters does DLDL (PASCAL) have?

DLDL (PASCAL) has 564M parameters. Finetunes 4 versions of VGG VGG 16: 138000000 VGG 19: 144000000 Total 2*(138000000+144000000)=564000000. That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.

Who created DLDL (PASCAL)?

DLDL (PASCAL) was published by University of Oxford, based in United Kingdom of Great Britain and Northern Ireland, categorised as academia.

When was DLDL (PASCAL) released?

DLDL (PASCAL) was published in November 2016. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

What is DLDL (PASCAL) used for?

DLDL (PASCAL) works in Vision, and is recorded as handling image classification. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download DLDL (PASCAL)?

The weights for DLDL (PASCAL) are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

Source

Original publication

Record last updated 11 February 2026

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.

All AI models