Reclaimed Node logoRECLAIMED NODE
← Back to Blog Homelab

Mac Studio vs. a 64GB NVIDIA CMP 170HX: Local AI Image Generation Head-to-Head

Oct 5, 2026 · 6 min read

Weathered green enamel sign on a brick wall reading Reclaimed Node - Second Life for Serious Hardware, Est. 2026, generated locally with Qwen Image 2512
Every image in this post came from the original 170HX run. The Mac Studio regenerated each one from the same prompt and seed.

In the last post I ran Qwen Image 2512 at full bf16 precision on an NVIDIA CMP 170HX unlocked to 64GB, a card originally sold for crypto mining. The whole 40GB model fit in VRAM with no quantization and no offloading. This time I put it up against a Mac Studio with an M5 Ultra and 256GB of unified memory, which shares one big pool of memory between its CPU and GPU. Same software, same model, same six prompts, same seeds.

The Setup

Mac Studio
NVIDIA CMP 170HX
Hardware
Apple M5 Ultra, 36-core CPU, 80-core GPU, 256GB unified memory
CMP 170HX, memory unlocked to 64GB, Ubuntu VM
Backend
PyTorch MPS (Metal)
CUDA
Software
InvokeAI 6.14
InvokeAI 6.14
Model
Qwen Image 2512, full bf16 (~40GB)
Qwen Image 2512, full bf16 (~40GB)
Text encoder
Qwen2.5-VL, bf16
Qwen2.5-VL, fp8 scaled
VAE
Qwen Image VAE
Qwen Image VAE
Speed LoRA
Qwen Image Lightning 4-step V2.0, weight 1
Qwen Image Lightning 4-step V2.0, weight 1
Sampling
4 steps, CFG 1
4 steps, CFG 1
Resolution
1680×944
1680×944
Model memory in use
54.5 GB
55.3 GiB

Both machines kept the entire model loaded between images. The Mac also had a 27B-parameter language model loaded in llama.cpp the whole time.

13.4 s vs 15.7 s

Average per image at 1680×944. Mac Studio ~15% faster.

9.2 s vs 14.3 s

Average per image at 1360×768. Mac Studio ~35% faster.

~17 s vs 41–109 s

First image, including model load.

Time Per Image

Seconds per image. Lower is better.

Mac Studio (M5 Ultra) CMP 170HX
1680×944
Enamel sign
12.38 s
15.31 s
GPU macro
14.57 s
15.68 s
Data center aisle
12.95 s
15.66 s
Night market
12.82 s
16.19 s
Mountain lake
14.76 s
15.63 s
Spacecraft sunrise
12.70 s
15.68 s
1360×768
Lighthouse
10.53 s
14.09 s

The Results

01Text Rendering: Enamel Sign

Green enamel sign on a brick wall reading Reclaimed Node - Second Life for Serious Hardware, Est. 2026
Seed 1131013313 Mac Studio 12.38 s CMP 170HX 15.31 s
Prompt

A vintage-style enamel sign mounted on a brick wall in warm late-afternoon light, hand-painted lettering reading "RECLAIMED NODE - SECOND LIFE FOR SERIOUS HARDWARE" in cream and deep green, with a smaller line beneath reading "Est. 2026". Slight weathering and chipped paint at the edges, soft rim light from a low sun, shallow depth of field, photorealistic, shot on an 85mm lens.

02Fine Detail: GPU Macro

Macro view of a gold-plated GPU heatsink with copper fins on a circuit board, lit from the side
Seed 47068114 Mac Studio 14.57 s CMP 170HX 15.68 s
Prompt

Extreme macro photograph of a data center GPU with a large gold-plated heatsink and dense rows of copper fins, lit by a single hard side light that rakes across the metal and reveals fine machining marks. Deep shadows, tiny dust particles catching the light, cool blue accent glow from an LED on the board, shallow depth of field, ultra-detailed, photorealistic.

03Perspective: Data Center Aisle

Symmetrical data center aisle of black server racks with green and blue status lights and an open door glowing amber at the far end
Seed 168100887 Mac Studio 12.95 s CMP 170HX 15.66 s
Prompt

A vast, quiet data center aisle stretching into the distance, rows of black server racks with thousands of tiny green and blue status lights, cold blue light from the ceiling and warm amber light spilling from an open door at the far end, light haze in the air, polished concrete floor with soft reflections. Symmetrical one-point perspective, cinematic, photorealistic, high dynamic range.

04Crowds: Night Market

Crowded rainy night market street with umbrellas, steaming food stalls, and neon signs
Seed 224642412 Mac Studio 12.82 s CMP 170HX 16.19 s
Prompt

A bustling night market in a rain-soaked neon city street, packed with hundreds of people under colorful umbrellas, food stalls with steam rising, hand-lettered signs in several languages, reflections of pink and teal neon in puddles, layers of depth from foreground to distant skyscrapers. Cinematic, richly detailed, photorealistic, shot on a 35mm lens.

05Reflections: Mountain Lake

Still mountain lake mirroring snow-capped peaks lit orange at sunrise, with mist and a wooden dock in the foreground
Seed 23128777 Mac Studio 14.76 s CMP 170HX 15.63 s
Prompt

A mountain lake at sunrise with perfectly still water mirroring snow-capped peaks, warm orange light striking the highest summits while the valley below stays in cool blue shadow, thin mist hovering over the water, a lone wooden dock in the foreground. Crisp detail, high dynamic range, photorealistic, landscape photography.

06Difficult Lighting: Spacecraft Sunrise

Research spacecraft with solar panels above the curved horizon of a blue planet at orbital sunrise, with lens flare and a dense star field
Seed 2245284484 Mac Studio 12.70 s CMP 170HX 15.68 s
Prompt

A lone research spacecraft with solar panels and brushed metal hull drifting above the curved limb of a blue ocean planet, at the moment of orbital sunrise. The sun is a blinding white point just cresting the horizon, throwing a thin band of glowing orange and gold along the atmosphere's edge and long lens flares across the frame. Hard directional sunlight rakes across the spacecraft, lighting its panels and edges with brilliant rim light while the shadowed side falls into deep blue-black, faintly lit by earthshine reflected from the planet below. Swirling white cloud systems and city lights on the dark side of the planet, thin blue atmospheric haze glowing at the horizon. Dense star field in the black sky above. Cinematic, high dynamic range, extreme contrast, photorealistic, shot on a 35mm lens.

07Lower Resolution: Lighthouse After a Storm

Stone cottage on a rocky coastline at golden hour with sun rays breaking through storm clouds and tidal pools reflecting the sky
Seed 1668950617 Mac Studio 10.53 s CMP 170HX 14.09 s
Prompt

A lone lighthouse keeper's cottage on a rocky coastline at golden hour, just after a storm. Low sun breaking through a gap in dark, heavy clouds, sending sharp god rays across the scene. Warm amber light rakes across wet stone and grass, creating long soft shadows and glowing rim light on the edges of the roof and rocks. Puddles and tidal pools on the rocks reflect the fiery orange and deep magenta sky. Cool blue shadows in the foreground contrast with the warm highlights. Atmospheric mist rising off the ocean, light scattering through it. Cinematic composition, high dynamic range, rich contrast, photorealistic, shot on a 50mm lens.

macOS GPU History window for the Apple M5 Ultra showing GPU usage pinned near 100% during two benchmark runs
The M5 Ultra’s GPU during a benchmark run, pinned for the entire set of images.

Where the Time Goes

One 1680×944 image on the Mac Studio:

Prompt encoding
0.35 s
4 denoising steps (~1.95 s per step)
7.8 s
LoRA and setup around denoising
2.2 s
VAE decode to the final image
2.3 s
Total
~12.7 s

What I Learned

01The Mac wins every image

At full resolution the M5 Ultra was faster on all six prompts, about 15% on average. Drop the resolution to 1360×768 and the gap widens to about 35%: the Mac scales down the way you’d expect, while the 170HX barely gets faster, which suggests a large fixed cost per image on that card.

02Loading is where the gap is biggest

The Mac loads the 55GB model in about 17 seconds. The 170HX took up to 109 seconds, because the weights come off a virtual disk and cross the card’s narrow PCIe link. You only pay this on the first image or when switching models, but you notice it every time.

03Check your model cache

Out of the box, Invoke on the Mac set aside only 32GB for cached models, less than the model needs. It reloaded the entire model before every image, at about 22 seconds each. One line in invokeai.yaml (max_cache_ram_gb: 128) kept everything loaded and brought it down to about 13 seconds.

04Restart Invoke now and then on a Mac

After about ten hours of running, large images crept up to about 25 seconds while small ones were unaffected. The GPU was still fast; the big memory operations around it had slowed down. A restart brought it straight back to about 13 seconds.

Times are per image, measured from Invoke’s queue with the model already loaded and Invoke’s result cache disabled, after one untimed warm-up image. The 170HX used an fp8 text encoder and the Mac a bf16 one, which affects only the sub-second prompt-encoding step. Images shown are the originals from the 170HX run; the Mac’s versions were visually near-identical.

Two very different routes to the same place: a mining card nobody wanted, rescued with a firmware unlock, and a desktop with enough unified memory to treat a 40GB model as ordinary. The Mac Studio is faster, loads quicker and does it while running a language model on the side. The 170HX costs a fraction as much and does the job at full precision. Both run a frontier-class image model at home with no cloud and no compromises.

Next: I put four of these cards in one server for 256GB of VRAM. See the quad CMP 170HX build and first LLM benchmarks.

Building Something Like This?

Tell us what you are trying to run — local models, passthrough labs, bulk storage — and we will tell you honestly what it takes and what we have on the bench.

Get Sourcing Help
More Posts Shop Inventory