SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2408.11049 · 2024

MagicDec: Breaking the Latency-Throughput Tradeoff for Long Context Generation with Speculative Decoding

arXiv · PDF · Open in the Atlas

Code that ran

4 of 5 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 3 failures of our instrument; 1 no recorded run); 1 of them was also run on the host, not in the jail, by the peer-agreement lane that set its level, measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2408.11049.json · badge: /api/ran/2408.11049.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
infini-ai-lab/magicdec canonical (classifier label)
classifier basis: census
4 ran · 1 no recorded run
FunctionStatusWhere it lives
repeat_kv Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
infini-ai-lab/magicdec/Engine/utils.py
code served (permissive licence) · get_code("3c76e52815c5401d")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 4, dtype family float, shape (2, 6, 5, 4) as a later run recorded it. The image it ran in was not recorded. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-04T23:23:17Z). If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small tensor with shape (batch, num_key_value_heads, seqlen, head_dim)
# Using distinct sizes: batch=2, num_key_value_heads=3, seqlen=5, head_dim=4
hidden_states = torch.randn(2, 3, 5, 4)

# n_rep is the repetition factor
n_rep = 2

ARGS = {
    "hidden_states": hidden_states,
    "n_rep": n_rep,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 4, dtype family float, shape (2, 6, 5, 4) — shape (2, 6, 5, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
dynamically_quantize_per_channel Ran
ran; our contract draft was wrong, not the code
infini-ai-lab/magicdec/Engine/quantize.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from amarka8/SOCKET, is pointer only (licence: NONE) · get_code("55ec8fea60e1d7e7")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from amarka8/SOCKET, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

find_multiple Ran
ran, honoured the contract we drafted
infini-ai-lab/magicdec/Engine/SnapKV/model.py
code served (permissive licence) · get_code("aba86ffbed27ba65")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

def find_multiple(n: int, k: int) -> int:
    if n % k == 0:
        return n
    return n + k - n % k

# The function takes two integers and returns an integer
# It finds the next multiple of k that is >= n
ARGS = {
    "n": 17,
    "k": 5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
unrepeat_kv Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
infini-ai-lab/magicdec/Engine/utils.py
code served (permissive licence) · get_code("ed50d08882ec36a7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 4, dtype family float, shape (2, 2, 4, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a repeated key-value tensor
# Shape: (batch, num_attention_heads, seqlen, head_dim)
# We'll use batch=2, num_attention_heads=8, seqlen=4, head_dim=3
batch = 2
num_attention_heads = 8
seqlen = 4
head_dim = 3
n_rep = 4  # number of repetitions (so num_key_value_heads = 8 // 4 = 2)

repeated_states = torch.randn(batch, num_attention_heads, seqlen, head_dim)

ARGS = {
    "repeated_states": repeated_states,
    "n_rep": n_rep,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 4, dtype family float, shape (2, 2, 4, 3) — shape (2, 2, 4, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
init_dist Not yet run
unverified (harvested, no recorded run)
infini-ai-lab/magicdec/Engine/tp.py
code served (permissive licence) · get_code("e4f851f13f6f224e")
How it ran

no invocation recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 3 the same month, 3 later) and 2 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
repeat_kv 243 other papers
105 dated earlier, 8 the same month, 126 later, 2 same year, order not determinable, 2 undated · 22 only as an assembly brick
1901.02860 (2019-01)
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
2203.08913 (2022-03)
Memorizing Transformers
2305.17888 (2023-05)
LLM-QAT: Data-Free Quantization Aware Training for Large Language Models
and 240 more not named here
find_multiple 9 other papers
5 dated earlier, 1 the same month, 3 later
2302.13971 (2023-02)
LLaMA: Open and Efficient Foundation Language Models
2310.06434 (2023-10)
Whispering LLaMA: A Cross-Modal Generative Error Correction Framework for Speech Recognition
2310.08278 (2023-10)
Lag-Llama: Towards Foundation Models for Probabilistic Time Series Forecasting
and 6 more not named here
dynamically_quantize_per_channel 4 other papers
1 dated earlier, 1 the same month, 2 later
2308.02560 (2023-08)
From Discrete Tokens to High-Fidelity Audio Using Multi-Band Diffusion
2408.07092 (2024-08)
Post-Training Sparse Attention with Double Sparsity
2511.00351 (2025-11)
Reject Only Critical Tokens: Pivot-Aware Speculative Decoding
and 1 more not named here

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Large Language Models (LLMs) have become more prevalent in long-context applications such as interactive chatbots, document analysis, and agent workflows, but it is challenging to serve long-context requests with low latency and high throughput. Speculative decoding (SD) is a widely used technique to reduce latency losslessly, but the conventional wisdom suggests that its efficacy is limited to small batch sizes. In MagicDec, we show that surprisingly SD can achieve speedup even for a high throughput inference regime for moderate to long sequences. More interestingly, an intelligent drafting strategy can achieve better speedup with increasing batch size based on our rigorous analysis. MagicDec first identifies the bottleneck shifts with increasing batch size and sequence length, and uses these insights to deploy SD more effectively for high throughput inference. We leverage draft model with sparse KV cache to address the KV bottleneck, which scales with both sequence length and batch size. Additionally, we propose a theoretical model to select the optimal drafting strategy for maximum speedup. Our work highlights the broad applicability of speculative decoding in long-context serving, as it can enhance throughput and reduce latency without compromising accuracy. For moderate to long sequences, we demonstrate up to 2.51x speedup for Llama3.1-8B when serving batch sizes ranging from 32 to 256 on various types of hardware and tasks.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2408.11049")
get_code_for_paper("2408.11049")
have("2408.11049")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2408.11049.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2408.11049.svg)](https://syntology.ai/paper/2408.11049)

Connect an agent — have() is free.