{"id":21256,"date":"2026-01-27T04:16:32","date_gmt":"2026-01-26T19:16:32","guid":{"rendered":"https:\/\/aireviewirush.com\/?p=21256"},"modified":"2026-01-27T04:16:32","modified_gmt":"2026-01-26T19:16:32","slug":"nvidia-dgx-spark-thermal-check-how-oem-cooling-designs-stack-up","status":"publish","type":"post","link":"https:\/\/aireviewirush.com\/?p=21256","title":{"rendered":"NVIDIA DGX Spark Thermal Check: How OEM Cooling Designs Stack Up"},"content":{"rendered":"<p> <br \/>\n<\/p>\n<div>\n<p>After a protracted wait, the <a href=\"https:\/\/www.storagereview.com\/review\/nvidia-dgx-spark-review-the-ai-appliance-bringing-datacenter-capabilities-to-desktops\" target=\"_blank\" rel=\"noopener\">NVIDIA DGX Spark<\/a> was launched final 12 months, with the hope of fostering a brand new wave of AI engineers and researchers. At StorageReview, we\u2019ve acquired a number of DGX Spark techniques from varied companions for analysis, and we\u2019re creating some thrilling cluster tasks. Nonetheless, the thermal and energy traits throughout these totally different implementations proved too attention-grabbing to maintain below wraps whereas we end our full evaluations.<\/p>\n<p><img fetchpriority=\"high\" decoding=\"async\" class=\"aligncenter wp-image-152369 size-full\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-CES-Sparks-1.jpg\" alt=\"NVIDIA DGX Spark Thermal\" width=\"1500\" height=\"1231\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-CES-Sparks-1.jpg 1500w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-CES-Sparks-1-300x246.jpg 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-CES-Sparks-1-1024x840.jpg 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-CES-Sparks-1-768x630.jpg 768w\" sizes=\"(max-width: 620px) 100vw, 2000px\"\/><\/p>\n<p>What makes this comparability notably compelling is the chance to watch how totally different OEM companions have interpreted NVIDIA\u2019s reference design. Whereas the core electronics and motherboard stay fixed throughout all implementations, the thermal administration options, chassis designs, and airflow methods range significantly. These engineering choices manifest as measurable variations in working temperatures, and understanding these variations offers priceless perception for finish customers out there for some DGX Sparks.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-152386\" alt=\"\" width=\"1500\" height=\"913\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-Spark-Group.jpg 1500w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-Spark-Group-300x183.jpg 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-Spark-Group-1024x623.jpg 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-Spark-Group-768x467.jpg 768w\" data-lazy-sizes=\"(max-width: 620px) 100vw, 2000px\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-Spark-Group.jpg\"><noscript><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-152386\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-Spark-Group.jpg\" alt=\"\" width=\"1500\" height=\"913\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-Spark-Group.jpg 1500w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-Spark-Group-300x183.jpg 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-Spark-Group-1024x623.jpg 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/StorageReview-NVIDIA-Spark-Group-768x467.jpg 768w\" sizes=\"auto, (max-width: 620px) 100vw, 2000px\"><\/noscript><\/p>\n<p>On this evaluation, we current a side-by-side comparability of thermal and energy efficiency for 5 DGX Spark techniques: the <a href=\"https:\/\/www.storagereview.com\/review\/nvidia-dgx-spark-review-the-ai-appliance-bringing-datacenter-capabilities-to-desktops\" target=\"_blank\" rel=\"noopener\">NVIDIA Founders Version<\/a>, Gigabyte, Dell, Acer, and ASUS.<\/p>\n<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_53 counter-hierarchy ez-toc-counter ez-toc-grey ez-toc-container-direction\">\n<div class=\"ez-toc-title-container\">\n<p class=\"ez-toc-title \" >Table of Contents<\/p>\n<span class=\"ez-toc-title-toggle\"><a href=\"#\" class=\"ez-toc-pull-right ez-toc-btn ez-toc-btn-xs ez-toc-btn-default ez-toc-toggle\" aria-label=\"Toggle Table of Content\" role=\"button\"><label for=\"item-69e7e634e3eb8\" ><span class=\"\"><span style=\"display:none;\">Toggle<\/span><span class=\"ez-toc-icon-toggle-span\"><svg style=\"fill: #999;color:#999\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" class=\"list-377408\" width=\"20px\" height=\"20px\" viewBox=\"0 0 24 24\" fill=\"none\"><path d=\"M6 6H4v2h2V6zm14 0H8v2h12V6zM4 11h2v2H4v-2zm16 0H8v2h12v-2zM4 16h2v2H4v-2zm16 0H8v2h12v-2z\" fill=\"currentColor\"><\/path><\/svg><svg style=\"fill: #999;color:#999\" class=\"arrow-unsorted-368013\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" width=\"10px\" height=\"10px\" viewBox=\"0 0 24 24\" version=\"1.2\" baseProfile=\"tiny\"><path d=\"M18.2 9.3l-6.2-6.3-6.2 6.3c-.2.2-.3.4-.3.7s.1.5.3.7c.2.2.4.3.7.3h11c.3 0 .5-.1.7-.3.2-.2.3-.5.3-.7s-.1-.5-.3-.7zM5.8 14.7l6.2 6.3 6.2-6.3c.2-.2.3-.5.3-.7s-.1-.5-.3-.7c-.2-.2-.4-.3-.7-.3h-11c-.3 0-.5.1-.7.3-.2.2-.3.5-.3.7s.1.5.3.7z\"\/><\/svg><\/span><\/span><\/label><input aria-label=\"Toggle\" aria-label=\"item-69e7e634e3eb8\"  type=\"checkbox\" id=\"item-69e7e634e3eb8\"><\/a><\/span><\/div>\n<nav><ul class='ez-toc-list ez-toc-list-level-1 ' ><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/aireviewirush.com\/?p=21256\/#NVIDIA_DGX_Spark_Thermal_Check_Methodology\" title=\"NVIDIA DGX Spark Thermal Check Methodology\">NVIDIA DGX Spark Thermal Check Methodology<\/a><ul class='ez-toc-list-level-3'><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/aireviewirush.com\/?p=21256\/#Chart_Alignment_and_Environmental_Controls\" title=\"Chart Alignment and Environmental Controls\">Chart Alignment and Environmental Controls<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/aireviewirush.com\/?p=21256\/#Information_Assortment\" title=\"Information Assortment\">Information Assortment<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/aireviewirush.com\/?p=21256\/#NVIDIA_DGX_Spark_Thermal_Evaluation\" title=\"NVIDIA DGX Spark Thermal Evaluation\">NVIDIA DGX Spark Thermal Evaluation<\/a><ul class='ez-toc-list-level-3'><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/aireviewirush.com\/?p=21256\/#CPU_Temperature\" title=\"CPU Temperature\">CPU Temperature<\/a><ul class='ez-toc-list-level-4'><li class='ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/aireviewirush.com\/?p=21256\/#GPU_Temperature\" title=\"GPU Temperature\">GPU Temperature<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-7\" href=\"https:\/\/aireviewirush.com\/?p=21256\/#NVMe_Temperature\" title=\"NVMe Temperature\">NVMe Temperature<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-4'><a class=\"ez-toc-link ez-toc-heading-8\" href=\"https:\/\/aireviewirush.com\/?p=21256\/#NIC_Temperature\" title=\"NIC Temperature\">NIC Temperature<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-9\" href=\"https:\/\/aireviewirush.com\/?p=21256\/#GPU_Energy_Consumption\" title=\"GPU Energy Consumption\">GPU Energy Consumption<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-10\" href=\"https:\/\/aireviewirush.com\/?p=21256\/#The_Takeaway\" title=\"The Takeaway\">The Takeaway<\/a><\/li><\/ul><\/li><\/ul><\/nav><\/div>\n<h2><span class=\"ez-toc-section\" id=\"NVIDIA_DGX_Spark_Thermal_Check_Methodology\"><\/span>NVIDIA DGX Spark Thermal Check Methodology<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>The info introduced right here was gathered whereas working vLLM on-line serving benchmarks utilizing OpenAI\u2019s GPT-OSS-120B mannequin. The benchmark includes three distinct testing eventualities designed to train totally different elements of the inference pipeline.<\/p>\n<p>The equal state of affairs presents a balanced workload with 256 enter and 256 output tokens, leading to brief, symmetric calls for on each the prefill and decode phases. The prefill-heavy state of affairs shifts computational depth towards enter processing, with 4096 enter tokens producing 512 output tokens; this configuration saturates tensor cores throughout the immediate prefill\/encoding part whereas sustaining average decode calls for, as in code-completion eventualities.<\/p>\n<p>Conversely, the decode-heavy state of affairs reverses this relationship, utilizing 512 enter tokens to generate 4096 output tokens, just like asking a mannequin to write down an software utilizing a single immediate, creating sustained reminiscence bandwidth strain throughout the autoregressive token technology part.<\/p>\n<p>Every state of affairs was examined throughout batch sizes of 1, 2, 4, 8, 16, 32, 64, and 128, leading to 24 whole take a look at phases. A 30-second cooldown interval was additionally enforced between every subsequent stage. These cooldown intervals are seen within the charts as transient dips between take a look at phases, marked by vertical dashed traces that separate the shaded state of affairs areas.<\/p>\n<h3><span class=\"ez-toc-section\" id=\"Chart_Alignment_and_Environmental_Controls\"><\/span>Chart Alignment and Environmental Controls<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>The charts introduced use a stage-aligned time axis, that means the information from all techniques is synchronized by take a look at stage slightly than absolute wall-clock time. This system permits direct comparability of system conduct throughout equivalent workload phases, regardless that particular person techniques accomplished phases at barely totally different charges on account of their thermal throttling conduct and sustained efficiency traits.<\/p>\n<p>The 5 techniques have been examined concurrently below equivalent environmental circumstances, positioned adjoining to 1 one other in a managed room with a constant ambient temperature all through the take a look at. This parallel testing strategy ensures that any variations noticed are immediately attributable to the techniques\u2019 thermal designs slightly than environmental variables similar to ambient temperature drift, airflow patterns, or time-of-day variations.<\/p>\n<p>Moreover, all fashions had the most recent NVIDIA Ubuntu picture put in.<\/p>\n<h3><span class=\"ez-toc-section\" id=\"Information_Assortment\"><\/span>Information Assortment<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>System metrics have been collected at 1-second intervals utilizing a customized monitoring script that reads immediately from Linux kernel interfaces and nvidia-smi. We didn&#8217;t take direct measurements with thermal probes or any exterior energy monitoring.<\/p>\n<p>Word that NVMe drive temperatures can\u2019t be immediately in contrast apples-to-apples, as not all techniques have equivalent storage drives. Notably, the Asus system was the one unit configured with a 1TB Phison drive, the Dell featured a 4TB Phison drive, and all different techniques we examined featured a 4TB Samsung drive. As we undergo every evaluation, we are going to discover NVMe drive thermals in additional element, as even components similar to thermal bridging between the drive and the case backside aren\u2019t the identical throughout Spark fashions.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"NVIDIA_DGX_Spark_Thermal_Evaluation\"><\/span>NVIDIA DGX Spark Thermal Evaluation<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Earlier than diving into the person metrics, it\u2019s essential to know that the DGX Spark makes use of an built-in cooling system through which parts share thermal pathways. As seen within the Founders Version design, GPU warmth immediately influences the temperatures of adjoining parts, together with the CPU, NVMe storage, and community interfaces. This thermal interdependence signifies that GPU-intensive workloads create cascading thermal results all through the system, explaining why we see correlated temperature patterns throughout totally different parts.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-150040\" alt=\"\" width=\"1500\" height=\"1020\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22.jpg 1500w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22-300x204.jpg 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22-1024x696.jpg 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22-768x522.jpg 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22-1170x796.jpg 1170w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22-585x398.jpg 585w\" data-lazy-sizes=\"(max-width: 620px) 100vw, 2000px\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22.jpg\"><noscript><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-150040\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22.jpg\" alt=\"\" width=\"1500\" height=\"1020\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22.jpg 1500w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22-300x204.jpg 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22-1024x696.jpg 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22-768x522.jpg 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22-1170x796.jpg 1170w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2025\/10\/StorageReview-NVIDIA-DGX-Spark-22-585x398.jpg 585w\" sizes=\"auto, (max-width: 620px) 100vw, 2000px\"><\/noscript><\/p>\n<h3><span class=\"ez-toc-section\" id=\"CPU_Temperature\"><\/span>CPU Temperature<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>The CPU temperature information reveals essentially the most dramatic variations between implementations. The Acer system stands out instantly, peaking at simply 74.6\u00b0C throughout the demanding Prefill Heavy state of affairs whereas each different system climbed into the mid-to-upper 80s. That\u2019s a 10-14\u00b0C benefit over the competitors, a considerable thermal headroom that implies Acer invested considerably in cooling capability.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-152351 size-full\" alt=\"NVIDIA DGX Spark Thermal CPU Temps\" width=\"2560\" height=\"1280\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-scaled.png 2560w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-300x150.png 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-1024x512.png 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-768x384.png 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-1536x768.png 1536w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-2048x1024.png 2048w\" data-lazy-sizes=\"(max-width: 620px) 100vw, 2000px\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-scaled.png\"\/><noscript><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-152351 size-full\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-scaled.png\" alt=\"NVIDIA DGX Spark Thermal CPU Temps\" width=\"2560\" height=\"1280\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-scaled.png 2560w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-300x150.png 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-1024x512.png 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-768x384.png 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-1536x768.png 1536w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_cpu_temp_aligned-2048x1024.png 2048w\" sizes=\"auto, (max-width: 620px) 100vw, 2000px\"\/><\/noscript><\/p>\n<p>The Founders Version, Dell, and Gigabyte cluster along with practically equivalent thermal profiles, all reaching 87-88\u00b0C throughout peak masses. This similarity signifies that almost all companions have primarily matched NVIDIA\u2019s reference thermal design slightly than pushing past it. ASUS lands within the center, working a couple of levels cooler than this group however nonetheless properly above Acer\u2019s efficiency.<\/p>\n<p>The sawtooth sample all through Equal and Prefill Heavy displays the benchmark\u2019s cyclical nature, with clear restoration throughout cooldown intervals. Throughout Decode Heavy, temperatures stabilize at a extra sustained degree; this additionally hints on the memory-intensive nature of token technology and that the reminiscence configuration is a transparent bottleneck on this workload.<\/p>\n<h4><span class=\"ez-toc-section\" id=\"GPU_Temperature\"><\/span>GPU Temperature<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<p>GPU temperatures comply with an identical story, given the shared cooling infrastructure. Acer peaks at simply 68\u00b0C throughout Prefill Heavy, whereas the opposite 4 techniques all hit 80-82\u00b0C, a constant 12-14\u00b0C benefit that mirrors the CPU findings.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-152349 size-full\" alt=\"NVIDIA DGX Spark Thermal GPU Temps\" width=\"2560\" height=\"1280\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-scaled.png 2560w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-300x150.png 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-1024x512.png 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-768x384.png 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-1536x768.png 1536w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-2048x1024.png 2048w\" data-lazy-sizes=\"(max-width: 620px) 100vw, 2000px\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-scaled.png\"\/><noscript><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-152349 size-full\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-scaled.png\" alt=\"NVIDIA DGX Spark Thermal GPU Temps\" width=\"2560\" height=\"1280\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-scaled.png 2560w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-300x150.png 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-1024x512.png 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-768x384.png 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-1536x768.png 1536w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_gpu_temp_aligned-2048x1024.png 2048w\" sizes=\"auto, (max-width: 620px) 100vw, 2000px\"\/><\/noscript><\/p>\n<p>The tight clustering amongst Founders Version, Dell, Gigabyte, and ASUS suggests these implementations are approaching related thermal limits, possible the place the GPU\u2019s firmware-controlled thermal administration begins to intervene.<\/p>\n<h4><span class=\"ez-toc-section\" id=\"NVMe_Temperature\"><\/span>NVMe Temperature<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<p>Storage temperatures reveal how properly every chassis isolates the drive bay from compute warmth. Acer once more leads at 51.8\u00b0C peak versus 58-63\u00b0C for the others.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-152355 size-full\" alt=\"NVIDIA DGX Spark Thermal SSD Temps\" width=\"2560\" height=\"1280\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-scaled.png 2560w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-300x150.png 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-1024x512.png 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-768x384.png 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-1536x768.png 1536w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-2048x1024.png 2048w\" data-lazy-sizes=\"(max-width: 620px) 100vw, 2000px\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-scaled.png\"\/><noscript><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-152355 size-full\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-scaled.png\" alt=\"NVIDIA DGX Spark Thermal SSD Temps\" width=\"2560\" height=\"1280\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-scaled.png 2560w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-300x150.png 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-1024x512.png 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-768x384.png 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-1536x768.png 1536w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nvme_temp_aligned-2048x1024.png 2048w\" sizes=\"auto, (max-width: 620px) 100vw, 2000px\"\/><\/noscript><\/p>\n<p>The gradual temperature rise all through testing, impartial of workload spikes, signifies warmth soak from adjoining parts accumulating over time. For prolonged fine-tuning runs or heavy mannequin swapping, that is related for drive longevity and sustained write efficiency.<\/p>\n<p>As famous earlier, drive configurations different throughout techniques, so these comparisons carry some caveats. However Acer\u2019s constant benefit throughout each thermal metric factors to real cooling enhancements slightly than variations in drive efficiency.<\/p>\n<h4><span class=\"ez-toc-section\" id=\"NIC_Temperature\"><\/span>NIC Temperature<span class=\"ez-toc-section-end\"><\/span><\/h4>\n<p>Trying on the ConnectX-7 NIC temps, Acer peaks at 62\u00b0C whereas the Founders Version hits 75\u00b0C, a 13\u00b0C unfold according to the sample we\u2019ve seen throughout different parts. Gigabyte really runs cooler right here than Dell and ASUS, regardless of working hotter on the CPU, suggesting that warmth distribution varies throughout the chassis designs.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-152354\" alt=\"\" width=\"2560\" height=\"1280\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-scaled.png 2560w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-300x150.png 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-1024x512.png 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-768x384.png 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-1536x768.png 1536w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-2048x1024.png 2048w\" data-lazy-sizes=\"(max-width: 620px) 100vw, 2000px\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-scaled.png\"><noscript><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-152354\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-scaled.png\" alt=\"\" width=\"2560\" height=\"1280\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-scaled.png 2560w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-300x150.png 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-1024x512.png 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-768x384.png 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-1536x768.png 1536w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_nic_temp_aligned-2048x1024.png 2048w\" sizes=\"auto, (max-width: 620px) 100vw, 2000px\"><\/noscript><\/p>\n<h3><span class=\"ez-toc-section\" id=\"GPU_Energy_Consumption\"><\/span>GPU Energy Consumption<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>Lastly, we wished to test the GPU\u2019s energy consumption. Unsurprisingly, all 5 techniques exhibited remarkably related peak energy consumption throughout Prefill Heavy, starting from 69.3W (Acer) to 76.0W (Gigabyte).<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-152353\" alt=\"\" width=\"2560\" height=\"1280\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-scaled.png 2560w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-300x150.png 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-1024x512.png 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-768x384.png 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-1536x768.png 1536w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-2048x1024.png 2048w\" data-lazy-sizes=\"(max-width: 620px) 100vw, 2000px\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-scaled.png\"><noscript><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-152353\" src=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-scaled.png\" alt=\"\" width=\"2560\" height=\"1280\" srcset=\"https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-scaled.png 2560w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-300x150.png 300w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-1024x512.png 1024w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-768x384.png 768w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-1536x768.png 1536w, https:\/\/www.storagereview.com\/wp-content\/uploads\/2026\/01\/DGX_Spark_Partner_Model_Comparison_max_gpu_power_aligned-2048x1024.png 2048w\" sizes=\"auto, (max-width: 620px) 100vw, 2000px\"><\/noscript><\/p>\n<h3><span class=\"ez-toc-section\" id=\"The_Takeaway\"><\/span>The Takeaway<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>The info tells a transparent story: the Acer unit we had ran 10-15\u00b0C cooler than opponents throughout each metric, suggesting a basically superior cooling answer. Founders Version, Dell, and Gigabyte cluster collectively thermally, primarily matching NVIDIA\u2019s reference design. ASUS occupies the center floor. Energy consumption is uniform, confirming that the thermal variations are due solely to the cooling implementation. Gigabyte, nonetheless, did stand out from the pack. It provided probably the greatest cooling profiles with one of many highest GPU energy ranges, placing the most effective stability between cooling and energy.<\/p>\n<p>General, the associate fashions carry out very equally, and customers can not go mistaken with any of them. However curious readers ought to keep tuned for a complete evaluation of every system, protecting workload efficiency and everybody\u2019s favourite: teardowns.<\/p>\n<\/p><\/div>\n\n","protected":false},"excerpt":{"rendered":"<p>After a protracted wait, the NVIDIA DGX Spark was launched final 12 months, with the hope of fostering a brand new wave of AI engineers and researchers. At StorageReview, we\u2019ve acquired a number of DGX Spark techniques from varied companions for analysis, and we\u2019re creating some thrilling cluster tasks. Nonetheless, the thermal and energy traits [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":21258,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[5],"tags":[],"class_list":{"0":"post-21256","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-computer-hardware"},"_links":{"self":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/21256","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=21256"}],"version-history":[{"count":1,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/21256\/revisions"}],"predecessor-version":[{"id":21257,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/21256\/revisions\/21257"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/media\/21258"}],"wp:attachment":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=21256"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=21256"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=21256"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}