| # DeepSeek decode cost on B200 + 50 GB/s InfiniBand |
|
|
| 本文估算 `{deepseek-v4-pro, deepseek-v4-flash, deepseek-r1}` 在 **B200 GPU + 每卡 50 GB/s InfiniBand** 上、**decode throughput 打满**时的输出 token 成本。 |
|
|
| 结论先放前面。按常规 decode 平均 KV 长度约 5k,即 DeepSeek 官方 V3/R1 线上统计口径附近: |
|
|
| | Model | 主要瓶颈 | 饱和 decode 吞吐, 每张 B200 | Decode cost | |
| | --- | ---: | ---: | ---: | |
| | deepseek-v4-pro | MoE all-to-all network | 5,445 tok/s | $0.306 / M output tok | |
| | deepseek-v4-flash | MoE all-to-all network | 13,518 tok/s | $0.123 / M output tok | |
| | deepseek-r1 | MoE all-to-all network, compute close | 4,235-4,454 tok/s | $0.374-$0.394 / M output tok | |
|
|
| 如果 50 GB/s 是理论带宽、实际只到 30 GB/s,那么上面的 network-bound 成本约乘以 `50 / 30 = 1.667`: |
|
|
| | Model | 30 GB/s 实效带宽成本 | |
| | --- | ---: | |
| | deepseek-v4-pro | $0.510 / M output tok | |
| | deepseek-v4-flash | $0.206 / M output tok | |
| | deepseek-r1 | $0.624-$0.656 / M output tok | |
|
|
| 如果换成 **GB300 NVL72**,关键不是单卡算力,而是 MoE all-to-all 能否留在单个 72-GPU NVLink domain 里。若可以留在 rack 内,并先按同样 `$6 / GPU-hour` 归一化: |
|
|
| | Model, 5k KV on GB300 NVL72 | 主要瓶颈 | 吞吐, 每张 GB300 | Decode cost | |
| | --- | ---: | ---: | ---: | |
| | deepseek-v4-pro | compute | 12,605 tok/s | $0.132 / M output tok | |
| | deepseek-v4-flash | compute | 49,014 tok/s | $0.034 / M output tok | |
| | deepseek-r1 | compute | 6,158 tok/s | $0.271 / M output tok | |
|
|
| 如果实际 GB300 租金是 `P` 美元/GPU-hour,上表成本需要再乘以 `P / 6`。如果 MoE all-to-all 仍然跨 rack 走外部网络,成本会重新退回对应 effective network bandwidth 的 network-bound 结果。 |
|
|
| 如果是 1M context decode,V4 的 CSA/HCA attention 仍然可控,但 R1/MLA 会被 attention FLOPs 打穿: |
|
|
| | Model, 1M context | 主要瓶颈 | 吞吐, 每张 B200 | Decode cost | |
| | --- | ---: | ---: | ---: | |
| | deepseek-v4-pro | compute | 4,403 tok/s | $0.379 / M output tok | |
| | deepseek-v4-flash | compute | 11,818 tok/s | $0.141 / M output tok | |
| | deepseek-r1 | compute | 39.6 tok/s | $42.0 / M output tok | |
|
|
| ## 0. Sources used |
|
|
| 本估算主要参考: |
|
|
| - [DeepSeek_V4.pdf](DeepSeek_V4.pdf): V4-Pro/V4-Flash 参数、CSA/HCA attention、KV cache 与 inference framework。 |
| - [2412.19437v2.pdf](2412.19437v2.pdf): DeepSeek-V3/R1 的 MLA、MoE、decode deployment 结构。 |
| - [articles/02-27181462601-DeepSeek-V3-R1-推理系统概览.md](articles/02-27181462601-DeepSeek-V3-R1-推理系统概览.md): 官方线上系统摘要,包含 decode EP/DP、平均 KV 长度、H800 单卡吞吐。 |
| - [articles/05-29841050824-DeepSeek-V3-R1-推理效率分析(2)-DeepSeek-满血版逆向工程分析.md](articles/05-29841050824-DeepSeek-V3-R1-推理效率分析(2)-DeepSeek-满血版逆向工程分析.md): H800 decode timeline 与 peak throughput 逆向。 |
| - [articles/06-29540042383-DeepSeek-V3-R1-推理效率分析(3):Decode-配置泛化讨论.md](articles/06-29540042383-DeepSeek-V3-R1-推理效率分析(3):Decode-配置泛化讨论.md): decode 配置泛化公式。 |
| - NVIDIA B200 规格按官方公开口径:单卡约 180 GB HBM3e、7.7-8.0 TB/s HBM、FP8 sparse 9 PFLOP/s、BF16 sparse 4.5 PFLOP/s。本文按 dense 有效峰值取一半,即 FP8 4.5 PFLOP/s、BF16 2.25 PFLOP/s、FP4 9 PFLOP/s。 |
| - NVIDIA GB300 NVL72 规格按官方公开口径:单 rack 72 张 Blackwell Ultra GPU、130 TB/s NVLink、20 TB GPU memory、FP8/FP6 sparse 720 PFLOP/s、BF16 sparse 360 PFLOP/s、FP4 dense 1080 PFLOP/s。NVLink 5 按每 GPU 1,800 GB/s,72-GPU NVLink domain。参考 [NVIDIA GB300 NVL72](https://www.nvidia.com/en-us/data-center/gb300-nvl72/)、[NVIDIA NVL72 AI Factory hardware components](https://docs.nvidia.com/enterprise-reference-architectures/nvl72-ai-factory/latest/components.html) 与 [NVIDIA NVLink specifications](https://www.nvidia.com/en-us/data-center/nvlink/)。 |
|
|
| ## 1. Scope and assumptions |
|
|
| 只算 decode 输出 token 的 GPU 租金成本,不含: |
|
|
| - prefill 成本; |
| - KV cache on-disk storage / SSD / CPU / DRAM; |
| - IB 交换机、NVSwitch、机房、电力、运维; |
| - 空载、调度碎片、请求不均衡; |
| - MTP/speculative decoding 的接受率收益; |
| - 量化到 FP4/NVFP4 后可能带来的额外收益。 |
|
|
| 输入条件: |
|
|
| ```text |
| B200 price = $6 / GPU-hour |
| IB bandwidth = 50 GB/s per GPU |
| ``` |
|
|
| 这里的 `50 GB/s` 是 NDR 400 Gb/s InfiniBand 的理论线速口径。实际 serving 里能吃到多少,取决于消息大小、GPU Direct RDMA/NCCL/DeepEP 实现、PCIe/NUMA/NIC 绑定、交换机拓扑、拥塞和多租户干扰。一个实用的经验口径是: |
|
|
| | Scenario | Effective bandwidth | Multiplier vs 50 GB/s | |
| | --- | ---: | ---: | |
| | 理想大包 RDMA microbench | 45-49 GB/s | 1.02-1.11x | |
| | 调好的 GPU Direct / NCCL 大包通信 | 40-47 GB/s | 1.06-1.25x | |
| | 非拥塞、拓扑较好的真实训练或 serving 通信 | 35-43 GB/s | 1.16-1.43x | |
| | MoE all-to-all / dispatch-combine serving 路径 | 30-40 GB/s | 1.25-1.67x | |
| | 多租户、拥塞或拓扑/绑定未调好 | 20-30 GB/s | 1.67-2.50x | |
|
|
| 因此本文主表仍保留 `50 GB/s` 理论上限,同时用 `30 GB/s` 作为 conservative serving estimate;如果是专用且调优较好的集群,可以把 `40 GB/s` 作为 optimistic serving estimate。 |
|
|
| 成本换算: |
|
|
| ```text |
| cost_per_M = price_per_hour * 1e6 / (throughput_tok_s_per_gpu * 3600) |
| = 6 * 1e6 / (throughput * 3600) |
| = 1666.667 / throughput |
| ``` |
|
|
| 所以只要拿到每卡 decode 吞吐 `T`: |
|
|
| ```text |
| T = 5,000 tok/s -> $0.333 / M output tok |
| T = 10,000 tok/s -> $0.167 / M output tok |
| T = 20,000 tok/s -> $0.083 / M output tok |
| ``` |
|
|
| ## 2. Model constants |
|
|
| ### 2.1 DeepSeek-R1, approximated by V3 architecture |
|
|
| R1 推理结构按 V3/R1 同架构估算: |
|
|
| ```text |
| hidden size h = 7168 |
| layers = 61 |
| MoE layers = 58 |
| activated params = 37B |
| attention heads = 128 |
| MLA KV compressed dimension d_c = 512 |
| MLA RoPE dimension d_R = 64 |
| routed experts per token = 8 |
| shared expert = 1 |
| decode effective experts E = 9 |
| ``` |
|
|
| V3 PDF 的 decoding 部分把 shared expert 当作 routed expert 处理,所以每个 token 在 decode 中等价选择 9 个 experts。网络主表用 `58` 个 MoE 层;为了给调度和非 MoE 层边际开销留余量,也列一个 `61` 层保守口径。 |
|
|
| ### 2.2 DeepSeek-V4-Pro |
|
|
| 来自 V4 PDF: |
|
|
| ```text |
| total params = 1.6T |
| activated params = 49B |
| layers = 61 |
| hidden size h = 7168 |
| all Transformer blocks use MoE |
| routed experts = 384 |
| routed experts activated per token = 6 |
| shared expert = 1 |
| decode effective experts E = 7 |
| expert intermediate dimension = 3072 |
| |
| attention heads n_h = 128 |
| attention head dim c = 512 |
| CSA compression m = 4 |
| CSA sparse top-k = 1024 |
| HCA compression m' = 128 |
| sliding window n_win = 128 |
| CSA indexer heads n_I = 64 |
| CSA indexer dim c_I = 128 |
| ``` |
|
|
| V4-Pro 前 2 层是 HCA,后续 59 层 CSA/HCA interleaved。本文按: |
|
|
| ```text |
| CSA layers = 30 |
| HCA layers = 31 |
| ``` |
|
|
| 如果 interleaving 起始相反,attention FLOPs 会有很小变化;在 5k context 下主瓶颈是 MoE 网络,不影响主结论。 |
|
|
| ### 2.3 DeepSeek-V4-Flash |
|
|
| 来自 V4 PDF: |
|
|
| ```text |
| total params = 284B |
| activated params = 13B |
| layers = 43 |
| hidden size h = 4096 |
| all Transformer blocks use MoE |
| routed experts = 256 |
| routed experts activated per token = 6 |
| shared expert = 1 |
| decode effective experts E = 7 |
| expert intermediate dimension = 2048 |
| |
| attention heads n_h = 64 |
| attention head dim c = 512 |
| query compression d_c = 1024 |
| CSA compression m = 4 |
| CSA sparse top-k = 512 |
| HCA compression m' = 128 |
| sliding window n_win = 128 |
| CSA indexer heads n_I = 64 |
| CSA indexer dim c_I = 128 |
| ``` |
|
|
| V4-Flash 前 2 层是 pure sliding-window attention,后续 41 层 CSA/HCA interleaved。本文按: |
|
|
| ```text |
| SWA layers = 2 |
| CSA layers = 21 |
| HCA layers = 20 |
| ``` |
|
|
| ## 3. Network-bound estimate |
|
|
| 这是常规 5k KV decode 的主瓶颈。 |
|
|
| V4 PDF 的 EP 通信讨论给出一个很有用的简化:每个 token-expert pair 的 MoE compute 是 `6 * h * d` FLOPs,通信是 `3 * h` bytes。其中: |
|
|
| ```text |
| dispatch activation = h bytes # FP8 |
| combine output = 2h bytes # BF16 |
| total = 3h bytes |
| ``` |
|
|
| 于是每张 GPU 的 network-bound decode 吞吐近似为: |
|
|
| ```text |
| T_net = B / (L_moe * E * 3 * h) |
| |
| B = per-GPU IB bandwidth in bytes/s |
| L_moe = number of MoE layers |
| E = active experts including shared expert |
| h = hidden size |
| ``` |
|
|
| ### 3.1 V4-Pro |
|
|
| ```text |
| B = 50e9 bytes/s |
| L_moe = 61 |
| E = 6 routed + 1 shared = 7 |
| h = 7168 |
| |
| T_net = 50e9 / (61 * 7 * 3 * 7168) |
| = 5,445 tok/s/GPU |
| |
| cost = 1666.667 / 5445 |
| = $0.306 / M output tok |
| ``` |
|
|
| ### 3.2 V4-Flash |
|
|
| ```text |
| B = 50e9 bytes/s |
| L_moe = 43 |
| E = 7 |
| h = 4096 |
| |
| T_net = 50e9 / (43 * 7 * 3 * 4096) |
| = 13,518 tok/s/GPU |
| |
| cost = 1666.667 / 13518 |
| = $0.123 / M output tok |
| ``` |
|
|
| ### 3.3 DeepSeek-R1 |
|
|
| MoE-only 58 层口径: |
|
|
| ```text |
| B = 50e9 bytes/s |
| L_moe = 58 |
| E = 8 routed + 1 shared = 9 |
| h = 7168 |
| |
| T_net = 50e9 / (58 * 9 * 3 * 7168) |
| = 4,454 tok/s/GPU |
| |
| cost = 1666.667 / 4454 |
| = $0.374 / M output tok |
| ``` |
|
|
| 保守 61 层口径: |
|
|
| ```text |
| T_net = 50e9 / (61 * 9 * 3 * 7168) |
| = 4,235 tok/s/GPU |
| |
| cost = 1666.667 / 4235 |
| = $0.394 / M output tok |
| ``` |
|
|
| 因此 R1 在这个假设下给: |
|
|
| ```text |
| deepseek-r1 decode cost ~= $0.37-$0.39 / M output tok |
| ``` |
|
|
| ### 3.4 Bandwidth sensitivity |
|
|
| | Model | 50 GB/s cost | 40 GB/s cost | 35 GB/s cost | 30 GB/s cost | |
| | --- | ---: | ---: | ---: | ---: | |
| | V4-Pro | $0.306/M | $0.383/M | $0.437/M | $0.510/M | |
| | V4-Flash | $0.123/M | $0.154/M | $0.176/M | $0.206/M | |
| | R1, 58 MoE layers | $0.374/M | $0.468/M | $0.535/M | $0.624/M | |
| | R1, 61 conservative layers | $0.394/M | $0.492/M | $0.562/M | $0.656/M | |
|
|
| 这张表也是最重要的 sanity check:如果 IB 带宽不变,B200 的更高算力不会线性转化成 decode 降本,因为 MoE all-to-all 已经是主要瓶颈。 |
|
|
| ## 4. Compute-bound sanity check at 5k context |
|
|
| 为了确认上面的 network bound 是否真的卡住,需要估算 compute upper bound。这里取 B200 dense 峰值,并假设有效 MFU 为 30%: |
|
|
| ```text |
| FP8 dense peak = 4.5 PFLOP/s |
| BF16 dense peak = 2.25 PFLOP/s |
| FP4 dense peak = 9.0 PFLOP/s |
| MFU = 30% |
| |
| effective FP8 = 1.35 PFLOP/s |
| effective BF16 = 0.675 PFLOP/s |
| effective FP4 = 2.70 PFLOP/s |
| ``` |
|
|
| ### 4.1 R1 compute at KV length S |
|
|
| R1/V3 MLA 的每 token decode FLOPs 近似分成两部分。 |
|
|
| 第一部分是 active params 的 GEMM: |
|
|
| ```text |
| F_fixed = 2 * activated_params |
| = 2 * 37B |
| = 74 GFLOPs/token |
| ``` |
|
|
| 第二部分是 MLA attention 随 KV 长度增长的部分: |
|
|
| ```text |
| F_attn = L * 2 * S * n_h * (2 * d_c + d_R) |
| |
| L = 61 |
| S = average KV length |
| n_h = 128 |
| d_c = 512 |
| d_R = 64 |
| ``` |
|
|
| 当 `S = 4989`: |
|
|
| ```text |
| F_attn = 61 * 2 * 4989 * 128 * (2 * 512 + 64) |
| = 84.8 GFLOPs/token |
| |
| F_total = 74.0 + 84.8 |
| = 158.8 GFLOPs/token |
| ``` |
|
|
| 其中 fixed GEMM 近似按 FP8,attention core 近似按 BF16: |
|
|
| ```text |
| T_compute = 1 / (F_fixed / effective_FP8 + F_attn / effective_BF16) |
| = 1 / (74e9 / 1.35e15 + 84.8e9 / 0.675e15) |
| = 5,544 tok/s/GPU |
| ``` |
|
|
| 对比 network bound: |
|
|
| ```text |
| T_net = 4,235-4,454 tok/s/GPU |
| ``` |
|
|
| 所以 R1 在 5k KV 下是 network-bound,但 compute 已经比较接近。 |
|
|
| ### 4.2 V4 compute at KV length S |
|
|
| V4 的 attention 不再是 V3/R1 的全量 MLA,而是 CSA/HCA hybrid attention。简化估算如下。 |
|
|
| 固定 active-param GEMM: |
|
|
| ```text |
| V4-Pro: |
| F_fixed = 2 * 49B = 98 GFLOPs/token |
| |
| V4-Flash: |
| F_fixed = 2 * 13B = 26 GFLOPs/token |
| ``` |
|
|
| CSA core attention 每层: |
|
|
| ```text |
| F_CSA_core_per_layer = 4 * n_h * c * (min(S / m, top_k) + n_win) |
| ``` |
|
|
| 这里 `4` 来自 QK 和 PV 两次乘法,各自约 `2 * n_h * c * entries`。 |
|
|
| HCA core attention 每层: |
|
|
| ```text |
| F_HCA_core_per_layer = 4 * n_h * c * (S / m' + n_win) |
| ``` |
|
|
| CSA lightning indexer 需要扫描压缩后的 blocks,近似: |
|
|
| ```text |
| F_CSA_index_per_layer = 2 * n_I * c_I * (S / m) |
| ``` |
|
|
| 其中 indexer 按 FP4 估算,core attention 按 BF16 估算。 |
|
|
| #### V4-Pro, S = 4989 |
|
|
| ```text |
| CSA entries = min(4989 / 4, 1024) + 128 |
| = 1152 |
| |
| HCA entries = 4989 / 128 + 128 |
| = 167 |
| |
| F_fixed = 98.0 GFLOPs/token |
| F_core = 10.4 GFLOPs/token |
| F_index = 0.61 GFLOPs/token |
| F_total = 109.0 GFLOPs/token |
| |
| T_compute = 1 / (98e9 / 1.35e15 |
| + 10.4e9 / 0.675e15 |
| + 0.61e9 / 2.70e15) |
| = 11,331 tok/s/GPU |
| ``` |
|
|
| 对比: |
|
|
| ```text |
| T_net = 5,445 tok/s/GPU |
| ``` |
|
|
| 所以 V4-Pro 在 5k KV 下明显 network-bound。 |
|
|
| #### V4-Flash, S = 4989 |
|
|
| ```text |
| F_fixed = 26.0 GFLOPs/token |
| F_core = 2.23 GFLOPs/token |
| F_index = 0.43 GFLOPs/token |
| F_total = 28.66 GFLOPs/token |
| |
| T_compute = 44,002 tok/s/GPU |
| T_net = 13,518 tok/s/GPU |
| ``` |
|
|
| 所以 V4-Flash 在 5k KV 下也明显 network-bound。 |
|
|
| ### 4.3 5k KV final by min(compute, network) |
|
|
| | Model | T_compute | T_net | Final T | Cost | |
| | --- | ---: | ---: | ---: | ---: | |
| | V4-Pro | 11,331 | 5,445 | 5,445 | $0.306/M | |
| | V4-Flash | 44,002 | 13,518 | 13,518 | $0.123/M | |
| | R1, 58 MoE layers | 5,544 | 4,454 | 4,454 | $0.374/M | |
| | R1, 61 conservative layers | 5,544 | 4,235 | 4,235 | $0.394/M | |
|
|
| ## 5. 1M context estimate |
|
|
| 1M context 是 V4 论文重点场景。这里 V4 与 R1 的差异会非常大。 |
|
|
| ### 5.1 R1 at S = 1,000,000 |
|
|
| R1 的 MLA attention FLOPs 随 `S` 线性增长: |
|
|
| ```text |
| F_fixed = 74 GFLOPs/token |
| |
| F_attn = 61 * 2 * 1,000,000 * 128 * 1088 |
| = 16,990 GFLOPs/token |
| |
| F_total = 17,064 GFLOPs/token |
| ``` |
|
|
| 吞吐: |
|
|
| ```text |
| T_compute = 1 / (74e9 / 1.35e15 + 16990e9 / 0.675e15) |
| = 39.6 tok/s/GPU |
| |
| cost = 1666.667 / 39.6 |
| = $42.0 / M output tok |
| ``` |
|
|
| 这说明 R1/V3 结构并不适合原样跑 1M context decode。 |
|
|
| ### 5.2 V4-Pro at S = 1,000,000 |
|
|
| V4-Pro 的 CSA sparse top-k 会把 CSA core attention 限住: |
|
|
| ```text |
| CSA entries = min(1,000,000 / 4, 1024) + 128 |
| = 1152 |
| ``` |
|
|
| HCA 仍随 `S / 128` 增长,indexer 随 `S / 4` 增长: |
|
|
| ```text |
| F_fixed = 98.0 GFLOPs/token |
| F_core = 73.6 GFLOPs/token |
| F_index = 122.9 GFLOPs/token |
| F_total = 294.5 GFLOPs/token |
| |
| T_compute = 4,403 tok/s/GPU |
| T_net = 5,445 tok/s/GPU |
| Final T = 4,403 tok/s/GPU |
| |
| cost = 1666.667 / 4403 |
| = $0.379 / M output tok |
| ``` |
|
|
| ### 5.3 V4-Flash at S = 1,000,000 |
|
|
| ```text |
| F_fixed = 26.0 GFLOPs/token |
| F_core = 22.6 GFLOPs/token |
| F_index = 86.0 GFLOPs/token |
| F_total = 134.6 GFLOPs/token |
| |
| T_compute = 11,818 tok/s/GPU |
| T_net = 13,518 tok/s/GPU |
| Final T = 11,818 tok/s/GPU |
| |
| cost = 1666.667 / 11818 |
| = $0.141 / M output tok |
| ``` |
|
|
| ### 5.4 MFU sensitivity for V4 at 1M |
|
|
| 1M context 下 V4 更偏 compute-bound,因此 MFU 对成本有影响: |
|
|
| | Model | MFU 20% | MFU 30% | MFU 40% | |
| | --- | ---: | ---: | ---: | |
| | V4-Pro | $0.568/M | $0.379/M | $0.306/M, network capped | |
| | V4-Flash | $0.212/M | $0.141/M | $0.123/M, network capped | |
|
|
| ## 6. GB300 NVL72 estimate |
|
|
| GB300 NVL72 和前面的 B200 + IB 估计不是简单换一张更快的 GPU。它的核心变化是一个 rack 内有 72 张 Blackwell Ultra GPU,形成 72-GPU NVLink domain;如果 MoE all-to-all 可以留在 rack 内,通信带宽从外部 IB 的几十 GB/s 级别变成 NVLink 的 TB/s 级别。 |
|
|
| 这里先按 `$6 / GPU-hour` 做归一化,方便和 B200 表直接比较。实际 GB300 租金若为 `P` 美元/GPU-hour,则所有 GB300 成本乘以 `P / 6`。 |
|
|
| 官方 rack 级算力换算成每 GPU dense peak: |
|
|
| ```text |
| GB300 NVL72 rack = 72 GPUs |
| |
| FP8 sparse rack peak = 720 PFLOP/s -> dense ~= 360 PFLOP/s rack ~= 5.0 PFLOP/s/GPU |
| BF16 sparse rack peak = 360 PFLOP/s -> dense ~= 180 PFLOP/s rack ~= 2.5 PFLOP/s/GPU |
| FP4 dense rack peak = 1080 PFLOP/s -> dense ~= 15.0 PFLOP/s/GPU |
| |
| MFU = 30% |
| |
| effective FP8 = 1.50 PFLOP/s/GPU |
| effective BF16 = 0.75 PFLOP/s/GPU |
| effective FP4 = 4.50 PFLOP/s/GPU |
| ``` |
|
|
| NVLink 5 的 per-GPU bandwidth 按 1,800 GB/s 估算: |
|
|
| ```text |
| T_net_NVL = T_net_50GB * (1800 / 50) |
| ``` |
|
|
| 所以 5k KV 常规 decode 下的 NVLink network upper bound 已经远高于 compute bound: |
|
|
| | Model | T_compute on GB300 | T_net over NVLink | Final T | Cost at $6/GPU-hour | |
| | --- | ---: | ---: | ---: | ---: | |
| | V4-Pro | 12,605 | 196,020 | 12,605 | $0.132/M | |
| | V4-Flash | 49,014 | 486,648 | 49,014 | $0.034/M | |
| | R1 | 6,158 | 152,460-160,344 | 6,158 | $0.271/M | |
|
|
| 这意味着:如果 all-to-all 真的留在单个 NVL72 rack 内,GB300 会把常规 5k decode 从 network-bound 推到 compute-bound。 |
|
|
| 但这是一个很强的拓扑假设。DeepSeek V3/R1 官方 decode 配置是 EP144,而单个 NVL72 只有 72 张 GPU;R1 如果照搬 EP144 就会跨两个 rack,重新吃外部网络。要拿到上表结果,需要把专家并行、专家复制、路由 locality 或冗余方案改到单 rack 内可用。V4-Pro/Flash 的 72-GPU rack 内部署也需要保证 routed experts 和 shared expert 的通信不溢出到 rack 外。 |
|
|
| 如果 MoE all-to-all 仍然走外部网络,那么 GB300 的更高算力不会改变 network-bound 成本。比如 effective network bandwidth 只有 30 GB/s 时,5k KV 成本仍然是: |
|
|
| | Model | 30 GB/s external network cost at $6/GPU-hour | |
| | --- | ---: | |
| | V4-Pro | $0.510/M | |
| | V4-Flash | $0.206/M | |
| | R1 | $0.624-$0.656/M | |
|
|
| 1M context 下,如果 all-to-all 仍在 NVLink rack 内,V4 和 R1 仍然是 compute-bound: |
|
|
| | Model, 1M context on GB300 NVL72 | Final T | Cost at $6/GPU-hour | |
| | --- | ---: | ---: | |
| | V4-Pro | 5,242 tok/s | $0.318/M | |
| | V4-Flash | 15,020 tok/s | $0.111/M | |
| | R1 | 44.0 tok/s | $37.8/M | |
|
|
| 因此 GB300 NVL72 的主要价值不是把跨 rack MoE 网络瓶颈变快一点,而是尽量把 MoE serving 的通信问题变成 rack 内 NVLink 问题。一旦 workload spill 到外部网络,成本模型要优先看 external effective bandwidth,而不是 GB300 的 tensor core 峰值。 |
|
|
| ## 7. Interpretation |
|
|
| 几个直观结论: |
|
|
| 1. 在 5k KV 常规 decode 下,B200 的更强算力不是主要降本来源。50 GB/s IB 先成为 MoE all-to-all 的瓶颈。 |
| 2. V4-Pro 的 active params 是 49B,大于 R1/V3 的 37B,但 V4-Pro 每 token routed experts 从 8 降到 6;按 shared expert 也计入通信,V4-Pro 是 7 experts,R1 是 9 experts。所以在同样 50 GB/s IB 下,V4-Pro decode 成本略低于 R1。 |
| 3. V4-Flash 由于 hidden size、层数、active params 都明显更小,decode 成本约为 V4-Pro 的 40%,约为 R1 的三分之一。 |
| 4. 1M context 是 V4 架构优势最明显的地方。V4-Pro/Flash 的 CSA/HCA 把 long-context attention 控制住,而 R1/MLA 的 attention FLOPs 随 context 线性增长,成本会到几十美元每百万 output tokens 的量级。 |
| 5. 如果要让 B200 在常规 decode 上继续降本,关键不是再加 GPU peak FLOPs,而是降低 MoE 通信量、提高 effective IB bandwidth、做更好的 expert locality、用更 aggressive 的 quantization/combine 压缩,或者靠 MTP/speculative decoding 提高每次 forward 产出的有效 tokens。 |
| 6. GB300 NVL72 的关键分界线是 MoE all-to-all 是否在 72-GPU NVLink domain 内完成;留在 rack 内时 5k KV decode 变成 compute-bound,跨 rack 时仍然是 external network-bound。 |
|
|