Papers › dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Caching

dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Caching

17 May 2025arXiv:2506.06295archive 2025-07-28

Zhiyuan Liu, Yicun Yang, Yaojie Zhang, Junjie Chen, Chang Zou, Qingyuan Wei, Shaobo Wang, Linfeng Zhang

Autoregressive Models (ARMs) have long dominated the landscape of Large Language Models. Recently, a new paradigm has emerged in the form of diffusion-based Large Language Models (dLLMs), which generate text by iteratively denoising masked segments. This approach has shown significant advantages and potential. However, dLLMs suffer from high inference latency. Traditional ARM acceleration techniques, such as Key-Value caching, are incompatible with dLLMs due to their bidirectional attention mechanism. To address this specific challenge, our work begins with a key observation that dLLM inference involves a static prompt and a partially dynamic response, where most tokens remain stable across adjacent denoising steps. Based on this, we propose dLLM-Cache, a training-free adaptive caching framework that combines long-interval prompt caching with partial response updates guided by feature similarity. This design enables efficient reuse of intermediate computations without compromising model performance. Extensive experiments on representative dLLMs, including LLaDA 8B and Dream 7B, show that dLLM-Cache achieves up to 9.1 x speedup over standard inference without compromising output quality. Notably, our method brings dLLM inference latency close to that of ARMs under many settings. Codes are provided in the supplementary material and will be released publicly on GitHub.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 13 code samples harvested from 1 repository linked to this paper: 10 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 1 where Syntology's instrument failed) · 3 unverified. Where Syntology's instrument failed: 1 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: community (archive-listed): 13 samples from 1 repository, 10 ran (9 with no instrument failure; 1 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written on 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 1 of the 10 samples that ran has no recorded date.

maomaocun/dllm-cache mentioned on GitHubjaxApache-2.0Syntology: 13 samples, 10 ran (1 instrument-failed) · 3 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

13 samples harvested: 10 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 1 where Syntology's instrument failed) · 3 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · our draft was wrong
9ran
3unverified

Licence: 0 of the 13 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from maomaocun/dllm-cache. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

maomaocun/dllm-cache community (archive-listed) 13 samples, 10 ran (1 instrument-failed) · 3 unverified
add_gumbel_noise in 32 papers
census: 7 implementations from 12 papers, 2 distinct outputs on one shared input
demo_MMada_cache.py ran fingerprinted Apache-2.0 (permissive) · 008ab2cfccca5ac2 · report fixture: signature logits:Tensor/2/float, temperature:float/0/float · returned Tensor · level recorded 2026-09-25
add_gumbel_noise in 32 papers
census: 7 implementations from 12 papers, 2 distinct outputs on one shared input
demo_MMada_mmu_cache.py ran fingerprinted Apache-2.0 (permissive) · 22cb8c51f0a9038f · report fixture: signature logits:Tensor/2/float, temperature:float/0/float · returned Tensor · level recorded 2026-09-25
convert_to_float in 13 papers Dream_test_flops_script.py ran Apache-2.0 (permissive) · 3822a0903091ccc2 · report fixture: signature value:str/0/str · returned float · level recorded 2026-09-25
create_prompt in 50 papers LLama_test_flops_script.py ran Apache-2.0 (permissive) · 0976b0b29f8d1d22 · report fixture: signature tokenizer:DummyTokenizer/0/?, length:int/0/int · returned str · level recorded 2026-09-25
generate_with_cache demo_MMada_cache.py ran Apache-2.0 (permissive) · 89ffc142130a4ab0 · report fixture: signature model:callable/0/?, prompt:Tensor/2/int, steps:int/0/int, gen_length:int/0/int, block_length:int/0/int, temperature:float/0/float, cfg_scale:float/0/float, remasking:str/0/str, mask_id:int/0/int, attention_mask:Tensor/3/bool · returned Tensor · level recorded 2026-09-25
get_num_transfer_tokens in 22 papers
census: 5 implementations from 8 papers, 3 distinct outputs across 2 buckets, one shared input per bucket
demo_MMada_cache.py ran Apache-2.0 (permissive) · ab0375db4fa54d7c · report fixture: signature mask_index:Tensor/2/int, steps:int/0/int · returned Tensor · level recorded 2026-09-25
llada_attention_hook_for_cache dllm_cache/hooks/cache_hook_LLaDA_V.py ran Apache-2.0 (permissive) · c35f5a7a8c404548 · report fixture: signature self:callable/0/?, q_in_proj:Tensor/3/float, k_in_proj:Tensor/3/float, v_in_proj:Tensor/3/float, attention_bias:Tensor/4/float, layer_past:NoneType/0/?, use_cache:bool/0/bool, q_index:NoneType/0/? · returned tuple · level recorded 2026-09-25
mmu_generate_with_cache demo_MMada_mmu_cache.py ran Apache-2.0 (permissive) · 5e2dc62c98e4c717 · report fixture: signature model:callable/0/?, input_ids:Tensor/2/int, input_embeddings:NoneType/0/?, max_new_tokens:int/0/int, steps:int/0/int, block_length:int/0/int, temperature:float/0/float, top_k:NoneType/0/?, eot_token:NoneType/0/?, cfg_scale:float/0/float, remasking:str/0/str, mask_id:int/0/int, attention_mask:Tensor/3/bool · returned Tensor · level recorded 2026-09-25
refresh_index dllm_cache/hooks/cache_hook_Dream.py ran Apache-2.0 (permissive) · 17ea12936141556f · report fixture: signature new_features:Tensor/3/float, cached_features:Tensor/3/float, transfer_ratio:float/0/float, layer_id:int/0/int · returned Tensor · level recorded 2026-09-25
rotate_half in 424 papers
census: 24 implementations from 96 papers, 4 distinct outputs across 2 buckets, one shared input per bucket this sample: bucket 1 of 2, arg 1: rank 2, kind float, dtype float32
dllm_cache/hooks/cache_hook_LLaDA_V.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · e03d53ba9d4f9ae5 · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded, date not recorded
RoPe_forward dllm_cache/hooks/cache_hook_LLaDA.py unverified Apache-2.0 (permissive) · 2af6f8d97abc409f · report

Tasks

Denoising

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections