NVIDIA H200 141GBコスパ重視141 GB VRAM (+13)4800 GB/s (+4527)
A
Raises estimated decode speed by about 779%.182 tok/sデコード
Raises estimated decode speed by about 779%.
Adds memory headroom for longer context windows and future model growth.
〜$30,000 MSRP
Can CodeLlama 13B Instruct run on NVIDIA DGX Spark 128GB?
YES — With F16
B70Good
Estimated from fit model
CodeLlama 13B Instruct needs ~53.1 GB VRAM. NVIDIA DGX Spark 128GB has 0 MB. With F16 quantization, expect ~9 tok/s.
Runtime: OllamaCapacity: RoomyBandwidth: LowStack: BasicBottleneck: Balanced
Operating mode
Choose the run profile you care about
Interactive favors responsiveness, while light API and scale-out lean harder on serving readiness. The fit stays the same, but the recommendation lens changes.
Current mode
Balanced
Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.
CodeLlama 13B Instruct at Q4_K_M needs 21.3 GB — too much for NVIDIA DGX Spark 128GB (0.0 GB). Runs at F16 (53.1 GB) with maximum quality. 8 quantization levels fit.
Select quantization to explore
Q4_K_M (Medium quality) — 34.4 GB, 20.7 tok/s, Runs well
34.4 GB required108.8 GB available
32% VRAM used
Fit status
Runs well
Decode
20.7 tok/s
TTFT
9373 ms
Safe context
16K
Memory
34.4 GB / 108.8 GB
Memory breakdown
Weights7.9 GB
KV Cache12.2 GB
Runtime1.2 GB
Headroom13.1 GB
See how fast it feels
See how fast it feelsCodeLlama 13B Instruct on NVIDIA DGX Spark 128GB
1st promptCold start — includes initialization
>Write a Python function to sort a list
Here's a Python function that sorts a list using quicksort:
def quicksort(arr):
if len(arr) <= 1:
return arr
pivot = arr[len(arr) // 2]
left = [x for x in arr if x < pivot]
middle = [x for x in arr if x == pivot]
right = [x for x in arr if x > pivot]
return quicksort(left) + middle + quicksort(right)
# Example usage
numbers = [3, 6, 8, 10, 1, 2, 1]
print(quicksort(numbers)) # [1, 1, 2, 3, 6, 8, 10]
The algorithm picks a pivot, partitions the list into smaller, equal, and larger elements, then recursively sorts each partition and concatenates the results.
2nd promptWarm — model ready, no init overhead
>Now add type hints and a docstring
def quicksort(arr: list[int]) -> list[int]:
"""Sort a list using the quicksort algorithm.
Args:
arr: List of integers to sort.
Returns:
A new sorted list.
"""
if len(arr) <= 1:
return arr
pivot = arr[len(arr) // 2]
left = [x for x in arr if x < pivot]
middle = [x for x in arr if x == pivot]
right = [x for x in arr if x > pivot]
return quicksort(left) + middle + quicksort(right)
Added type hints for the input and return type, plus a Google-style docstring.
Estimated: 20.7 tok/s decode · 9.4s TTFT (warm) · 52 tok/s prefill
What limits this setup
This setup is broadly balanced for this model.
Shared-memory contention still exists
The OS, browser, and inference runtime all compete for the same physical memory pool, so real-world headroom is less forgiving than raw capacity suggests.
Best improvement path
Performance by workload
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Chat | F | Too heavy | 3.7 tok/s | 28402 ms | 4K |
| Coding | F | Too heavy | 3.7 tok/s | 52071 ms | 4K |
| Agentic Coding | F | Too heavy | 3.7 tok/s | 75739 ms | 4K |
| Reasoning | F | Too heavy | 3.7 tok/s | 61538 ms | 4K |
| RAG | F | Too heavy | 3.7 tok/s | 94674 ms | 4K |
Quantization options
How CodeLlama 13B Instruct (13B params) fits at each quantization level on NVIDIA DGX Spark 128GB (92.2 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 5.1 GB | Low | B64 |
Q3_K_S | 3 | 6.4 GB | Low | B64 |
NVFP4 | 4 | 7.3 GB | Medium | B64 |
Q4_K_M | 4 | 7.9 GB | Medium | B65 |
Q5_K_M | 5 | 9.4 GB | High | B65 |
Q6_K | 6 | 10.7 GB | High | B65 |
Q8_0 | 8 | 13.9 GB | Very High | B65 |
F16Best for your GPU | 16 | 26.7 GB | Maximum | B67 |
Get started
Copy-paste commands to run CodeLlama 13B Instruct on your machine.
Run
lms load CodeLlama-13b-Instruct-hf && lms server startアップグレードオプション
CodeLlama 13B Instructを快適に動かすハードウェア
NVIDIA H200 141GBコスパ重視NVIDIA H200 PCIe 141GBコスパ最良141 GB VRAM (+13)4800 GB/s (+4527)
A
Raises estimated decode speed by about 779%.182 tok/sデコード
Raises estimated decode speed by about 779%.
Adds memory headroom for longer context windows and future model growth.
〜$30,000 MSRP
NVIDIA B200 180GBNVIDIAアップグレード180 GB VRAM (+52)8000 GB/s (+7727)
A
Raises estimated decode speed by about 779%.182 tok/sデコード
Raises estimated decode speed by about 779%.
Adds memory headroom for longer context windows and future model growth.
〜$30,000 MSRP
Frequently asked questions
Can NVIDIA DGX Spark 128GB run CodeLlama 13B Instruct?
Yes, NVIDIA DGX Spark 128GB can run CodeLlama 13B Instruct at F16 quantization (Runs well). The recommended Q4_K_M requires 21.3 GB which exceeds available memory, but at F16 it needs only 53.1 GB. Expected decode speed: 8.6 tok/s.
How much VRAM does CodeLlama 13B Instruct need?
CodeLlama 13B Instruct (13B parameters) requires approximately 21.3 GB at Q4_K_M quantization. On NVIDIA DGX Spark 128GB, it fits at F16 using 53.1 GB.
What is the best quantization for CodeLlama 13B Instruct?
The recommended quantization is Q4_K_M, but on NVIDIA DGX Spark 128GB the best fitting quantization is F16, which uses 53.1 GB.
What speed will CodeLlama 13B Instruct run at on NVIDIA DGX Spark 128GB?
On NVIDIA DGX Spark 128GB, CodeLlama 13B Instruct achieves approximately 8.6 tokens per second decode speed with a time-to-first-token of 22499ms using F16 quantization.
Can NVIDIA DGX Spark 128GB run CodeLlama 13B Instruct for coding?
For coding workloads, CodeLlama 13B Instruct on NVIDIA DGX Spark 128GB receives a F grade with 3.7 tok/s and 4K context.
What context window can CodeLlama 13B Instruct use on NVIDIA DGX Spark 128GB?
On NVIDIA DGX Spark 128GB, CodeLlama 13B Instruct can safely use up to 16K tokens of context at F16 quantization. The model's official context limit is 16K, but available memory constrains the safe maximum.
Is unified memory on NVIDIA DGX Spark 128GB as fast as VRAM for CodeLlama 13B Instruct?
Not always. NVIDIA DGX Spark 128GB can often fit larger models thanks to unified memory, but a discrete GPU with dedicated high-bandwidth VRAM may still decode faster once the model fits. For this combination, the important distinction is capacity versus sustained throughput.
Embed this result▼
Paste this snippet into any page to show a live fit card.
<iframe src="https://willitrunai.com/embed/codellama-13b-instruct-on-dgx-spark-128gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>
Preview: