AWD-LSTM-DRILL + dynamic evaluation† (WT2) TPS calculator

Open weights IDIAP 34M parameters May 2019

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 · 1,084 tok/s

Fastest card

B200

99,654 tok/s · 180 GB

Which GPUs can run AWD-LSTM-DRILL + dynamic evaluation† (WT2)?

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
99,654 tok/s 59,792–159,446 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.7 GB	Q8_0	Comfortable
99,654 tok/s 59,792–159,446 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.7 GB	Q8_0	Comfortable
79,576 tok/s 47,746–127,322 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
79,576 tok/s 47,746–127,322 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.7 GB	Q8_0	Comfortable
63,642 tok/s 38,185–101,826 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
60,913 tok/s 36,548–97,462 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
60,913 tok/s 36,548–97,462 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.7 GB	Q8_0	Comfortable
58,298 tok/s 34,979–93,276 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.7 GB	Q8_0	Comfortable
51,739 tok/s 31,043–82,783 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
51,739 tok/s 31,043–82,783 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
51,739 tok/s 31,043–82,783 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.7 GB	Q8_0	Comfortable
49,080 tok/s 29,448–78,527 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
41,855 tok/s 25,113–66,967 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
41,855 tok/s 25,113–66,967 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.7 GB	Q8_0	Comfortable
41,855 tok/s 25,113–66,967 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
41,855 tok/s 25,113–66,967 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
41,855 tok/s 25,113–66,967 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.7 GB	Q8_0	Comfortable
31,869 tok/s 19,122–50,991 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
31,869 tok/s 19,122–50,991 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.7 GB	Q8_0	Comfortable
26,558 tok/s 15,935–42,492 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
25,991 tok/s 15,595–41,586 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.7 GB	Q8_0	Comfortable
25,412 tok/s 15,247–40,659 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.7 GB	Q8_0	Comfortable
25,412 tok/s 15,247–40,659 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.7 GB	Q8_0	Comfortable
25,412 tok/s 15,247–40,659 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.7 GB	Q8_0	Comfortable
25,412 tok/s 15,247–40,659 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	0.7 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: IDIAP
Organisation type: Academia
Country: Switzerland
Published: 14 May 2019
Authors: Nikolaos Pappas, James Henderson

What it does

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

Domain: Language
Task: Language modeling
Numerical format: FP32

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: 34M
Training data: 2,000,000 tokens
Epochs: 1,000

Training compute

The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.

Training compute: 4.1 × 10¹⁷ FLOP
How it was established: Operation counting

The training run

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

Wall-clock time: 29 hours

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)
Training code: Open source

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
Citations: 7
Benchmark data: AWD-LSTM-DRILL + dynamic evaluation† (WT2)

Sources

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

Reference: Deep Residual Output Layers for Neural Language Generation
Last updated: 11 February 2026

The extremes

The ten fastest GPUs that run AWD-LSTM-DRILL + dynamic evaluation† (WT2)

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 AWD-LSTM-DRILL + dynamic evaluation† (WT2)

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

The hardware side

Minimum card

Tesla C1080

Memory needed

0.7 GB

Fastest

99,654 tok/s

AWD-LSTM-DRILL + dynamic evaluation† (WT2) is small enough at 34M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

At the low end, a Tesla C1080 handles it — 4 GB, at Q8_0, for about 1,084 tokens per second.

At the other end, a B200 generates roughly 99,654 tokens per second on it, on the strength of 8,000 GB/s of memory bandwidth.

About this model

AWD-LSTM-DRILL + dynamic evaluation† (WT2) was published by IDIAP, in Switzerland, in May 2019. academia is the category the publisher falls under.

It works in Language, and is recorded as doing language modeling.

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.

How fast it runs, and why

The median result is around 2,798.3 tokens per second; 818 cards produce text faster than most people read it.

Every weight participates in every token here, so bandwidth is the whole story: the ranking below is effectively a ranking of memory throughput.

Memory here is estimated from size rather than computed from the architecture, which is not recorded for this model — the numbers are indicative rather than exact.

How it was trained

Producing it required around 4.1 × 10¹⁷ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

Around 2,000,000 tokens went into training it.

The reason it appears in this catalogue at all is sOTA improvement.

Step by step

How to choose a GPU for AWD-LSTM-DRILL + dynamic evaluation† (WT2)

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 AWD-LSTM-DRILL + dynamic evaluation† (WT2) — around 0.7 GB at Q8_0. Capacity is the gate — a card either holds it or it does not.
02

Decide how long your conversations run

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for AWD-LSTM-DRILL + dynamic evaluation† (WT2).
03

Set a quality floor

The quantisation column varies by card, because a bigger card holds a more accurate copy of AWD-LSTM-DRILL + dynamic evaluation† (WT2) — Q8_0 on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Sort by speed

The speed ordering for AWD-LSTM-DRILL + dynamic evaluation† (WT2) is effectively an ordering by memory bandwidth, which is why the B200 tops it at 99,654 tok/s.
05

Look at the headroom, not just the fit

A tight fit runs AWD-LSTM-DRILL + dynamic evaluation† (WT2) but leaves nothing spare for a longer conversation; comfortable has headroom. If you expect to grow the context, buy for comfortable.
06

Check the card from the other side

Following a card through to its own page shows every other model it can hold, which is the question that follows once AWD-LSTM-DRILL + dynamic evaluation† (WT2) is settled.

Answers

AWD-LSTM-DRILL + dynamic evaluation† (WT2) — common questions

Who created AWD-LSTM-DRILL + dynamic evaluation† (WT2)?

AWD-LSTM-DRILL + dynamic evaluation† (WT2) was published by IDIAP, based in Switzerland, categorised as academia.

When was AWD-LSTM-DRILL + dynamic evaluation† (WT2) released?

AWD-LSTM-DRILL + dynamic evaluation† (WT2) was published in May 2019. 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 AWD-LSTM-DRILL + dynamic evaluation† (WT2) used for?

AWD-LSTM-DRILL + dynamic evaluation† (WT2) works in Language, and is recorded as handling language modeling. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

Where can I download AWD-LSTM-DRILL + dynamic evaluation† (WT2)?

The weights for AWD-LSTM-DRILL + dynamic evaluation† (WT2) are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

How much compute was used to train AWD-LSTM-DRILL + dynamic evaluation† (WT2)?

Around 4.1 × 10¹⁷ FLOP. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.

Can I run AWD-LSTM-DRILL + dynamic evaluation† (WT2) 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 AWD-LSTM-DRILL + dynamic evaluation† (WT2) assume it is fully resident.

Would two GPUs run AWD-LSTM-DRILL + dynamic evaluation† (WT2) faster?

Two cards buy memory rather than speed. That matters for AWD-LSTM-DRILL + dynamic evaluation† (WT2) only if one card cannot hold it — 818 can, so a second adds little.

Why does the quantisation differ between cards for AWD-LSTM-DRILL + dynamic evaluation† (WT2)?

Because capacity varies, so does how hard AWD-LSTM-DRILL + dynamic evaluation† (WT2) has to be squeezed — 1 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these AWD-LSTM-DRILL + dynamic evaluation† (WT2) speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 59,792–159,446 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 AWD-LSTM-DRILL + dynamic evaluation† (WT2)?

The smallest card in our catalogue that holds AWD-LSTM-DRILL + dynamic evaluation† (WT2) is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.7 GB, and produces roughly 1,084 tokens per second. 818 cards in total can run it.

How fast is AWD-LSTM-DRILL + dynamic evaluation† (WT2) on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 99,654 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 818 of the cards that can run AWD-LSTM-DRILL + dynamic evaluation† (WT2) clear that.

How much VRAM does AWD-LSTM-DRILL + dynamic evaluation† (WT2) need?

About 0.7 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 AWD-LSTM-DRILL + dynamic evaluation† (WT2) on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 0.7 GB and generating roughly 18,561 tokens per second — a comfortable fit.

Can I run AWD-LSTM-DRILL + dynamic evaluation† (WT2) on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 0.7 GB and generating roughly 11,366 tokens per second — a comfortable fit.

Can I run AWD-LSTM-DRILL + dynamic evaluation† (WT2) on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 0.7 GB and generating roughly 14,076 tokens per second — a comfortable fit.

Can I run AWD-LSTM-DRILL + dynamic evaluation† (WT2) on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 0.7 GB and generating roughly 16,692 tokens per second — a comfortable fit.

Is AWD-LSTM-DRILL + dynamic evaluation† (WT2) open source?

Its weights are published, so AWD-LSTM-DRILL + dynamic evaluation† (WT2) 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 AWD-LSTM-DRILL + dynamic evaluation† (WT2) have?

AWD-LSTM-DRILL + dynamic evaluation† (WT2) has 34M parameters. 34M, Table 2. 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.

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