{"article":{"annotation":{"article_id":"nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026","available_at":"2026-10-15T19:01:06.304413Z","geography_revision":"ef6ee0010dda5fe7b240fd514749b43e214c57da73222d2e40ec83b276dd0977","id":"24200747d4eaa67025c14b642836c7e8b5a081ccec887cb248fd3eb9d2e9a9a3","input_sha256":"f1951e8e947f42b310ac8a41c7e409fae36d733af3c76caccdb2c14aa9abd70b","limitations":["Rule-derived subjects are candidates, not verified editorial labels or calibrated model probabilities.","Coverage is distinct from event location. Unknown country/city/coordinates remain absent; no automatic geocoding.","Article geometries keep their stated precision and methodology; reproducing them is not independent coordinate verification."],"method":"controlled-rules-v1","places":[],"rules_sha256":"fb5ad3c27c91cb840f7a0a05f7f64a5cb26ad0374409ff9457f277c7092e22b4","schema_version":1,"subjects":[{"concept_id":"medtop:13000000","confidence":null,"confidence_method":null,"evidence":[{"field":"category+topics","language":"en","note":"Preserved editorial category/topic; broad candidate only","source_urls":["https://developer.nvidia.com/blog/optimizing-agent-reliability-and-tool-calling-efficiency-for-chip-design-using-nvidia-nemo-gym/"]}],"method":"controlled-rules-v1","rule_id":"technology-broad","status":"candidate"}],"taxonomy_revision":"1f2db31d811fca07e99b16d402c10e2cc2965c9df1c7b41d9be348b40a24e7fb","taxonomy_version":"IPTC Media Topics 2026-Q2"},"api_url":"https://newsverum.com/api/v1/articles/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026?lang=fr","category":"technology","content":{"key_points":[],"paragraphs":["NVIDIA indique que Siemens EDA a utilisé NeMo Gym pour tester les fichiers Liberty générés par son agent IA Fuse EDA pendant la caractérisation de cellules standard. Un vérificateur contrôlait la couverture des cellules et conditions d’exploitation, la validité des fichiers et le respect des tolérances des valeurs de temporisation et de puissance. L’équipe a étudié les échecs puis modifié par itérations les compétences de l’agent et les descriptions d’outils Model Context Protocol.","Dans le benchmark de développement rapporté par NVIDIA, le score composite de vérification est passé d’environ 55 % à 95 %, les tours de l’agent de 155 à 53 et les jetons de sortie d’environ 81 000 à 29 000. Le dispositif couvrait 229 cellules et 36 conditions du kit de conception de procédé ouvert GF180MCU, tandis que les essais de réglage de l’agent portaient sur deux cellules ; l’équipe a mené environ 25 itérations. Ce sont des résultats de NVIDIA et Siemens sur des tâches représentatives, pas une mesure générale de la productivité en conception de puces."],"sections":[]},"corrections":[],"country":null,"data":{"blocks":[{"description":{"ar":"الدرجات التقريبية وأعداد خطوات سير العمل التي أبلغت عنها NVIDIA في معيار توصيف الخلايا القياسية.","de":"Von NVIDIA gemeldete ungefähre Werte und Arbeitsablaufzahlen für den Benchmark zur Standardzellencharakterisierung.","en":"Approximate score and workflow counts reported by NVIDIA for its standard-cell-characterization benchmark.","es":"Puntuaciones aproximadas y recuentos del flujo de trabajo comunicados por NVIDIA para su prueba de caracterización de celdas estándar.","fr":"Scores approximatifs et volumes de travail publiés par NVIDIA pour son benchmark de caractérisation de cellules standard.","pt":"Pontuações aproximadas e contagens do fluxo de trabalho divulgadas pela NVIDIA para o benchmark de caracterização de células padrão.","tr":"NVIDIA'nın standart hücre karakterizasyonu kıyaslamasındaki yaklaşık puan ve iş akışı sayıları."},"id":"siemens-eda-agent-evaluation","items":[{"basis":"observed","label":{"ar":"درجة التحقق المركبة قبل التحسين","de":"Zusammengesetzter Verifizierungswert zuvor","en":"Composite verification score, before","es":"Puntuación compuesta de verificación antes","fr":"Score composite de vérification avant optimisation","pt":"Pontuação composta de verificação antes","tr":"Bileşik doğrulama puanı, önce"},"measure":{"aggregation":"rate","frequency":"irregular","geography":"world","id":"siemens-fuse-agent-verification-score-before","periodRole":"announcement-date","unitCode":"percent"},"period":"2026-10-14","unit":"percent","value":55,"valueQualifier":"approximate"},{"basis":"observed","label":{"ar":"درجة التحقق المركبة بعد التحسين","de":"Zusammengesetzter Verifizierungswert danach","en":"Composite verification score, after","es":"Puntuación compuesta de verificación después","fr":"Score composite de vérification après optimisation","pt":"Pontuação composta de verificação depois","tr":"Bileşik doğrulama puanı, sonra"},"measure":{"aggregation":"rate","frequency":"irregular","geography":"world","id":"siemens-fuse-agent-verification-score-after","periodRole":"announcement-date","unitCode":"percent"},"period":"2026-10-14","unit":"percent","value":95,"valueQualifier":"approximate"},{"basis":"observed","label":{"ar":"عدد جولات الوكيل قبل التحسين","de":"Agentenschritte zuvor","en":"Agent turns, before","es":"Turnos del agente antes","fr":"Tours de l’agent avant optimisation","pt":"Turnos do agente antes","tr":"Ajan turları, önce"},"measure":{"aggregation":"unspecified","frequency":"irregular","geography":"world","id":"siemens-fuse-agent-turns-before","periodRole":"announcement-date","unitCode":"count"},"period":"2026-10-14","unit":"count","value":155},{"basis":"observed","label":{"ar":"عدد جولات الوكيل بعد التحسين","de":"Agentenschritte danach","en":"Agent turns, after","es":"Turnos del agente después","fr":"Tours de l’agent après optimisation","pt":"Turnos do agente depois","tr":"Ajan turları, sonra"},"measure":{"aggregation":"unspecified","frequency":"irregular","geography":"world","id":"siemens-fuse-agent-turns-after","periodRole":"announcement-date","unitCode":"count"},"period":"2026-10-14","unit":"count","value":53},{"basis":"observed","label":{"ar":"رموز المخرجات قبل التحسين","de":"Ausgabetoken zuvor","en":"Output tokens, before","es":"Tokens de salida antes","fr":"Jetons de sortie avant optimisation","pt":"Tokens de saída antes","tr":"Çıktı tokenları, önce"},"measure":{"aggregation":"unspecified","frequency":"irregular","geography":"world","id":"siemens-fuse-agent-output-tokens-before","periodRole":"announcement-date","unitCode":"count"},"period":"2026-10-14","unit":"count","value":81000,"valueQualifier":"approximate"},{"basis":"observed","label":{"ar":"رموز المخرجات بعد التحسين","de":"Ausgabetoken danach","en":"Output tokens, after","es":"Tokens de salida después","fr":"Jetons de sortie après optimisation","pt":"Tokens de saída depois","tr":"Çıktı tokenları, sonra"},"measure":{"aggregation":"unspecified","frequency":"irregular","geography":"world","id":"siemens-fuse-agent-output-tokens-after","periodRole":"announcement-date","unitCode":"count"},"period":"2026-10-14","unit":"count","value":29000,"valueQualifier":"approximate"}],"methodology":{"ar":"قيم ما قبل التحسين وما بعده الواردة في دراسة حالة Siemens EDA التي نشرتها NVIDIA في 14 أكتوبر 2026، بعد تعديل مهارات الوكيل وأوصاف أدوات MCP. استخدم المعيار 229 خلية قياسية و36 زاوية PVT من GF180MCU PDK؛ وقيمت جولات ضبط الوكيل خليتين. النتائج أرقام معيار تطوير أبلغ عنها المورد لمهام تمثيلية.","de":"Vorher-Nachher-Werte nach der Überarbeitung von Agentenfähigkeiten und MCP-Werkzeugbeschreibungen aus der Siemens-EDA-Fallstudie von NVIDIA vom 14. Oktober 2026. Der Benchmark nutzte 229 Standardzellen und 36 PVT-Ecken des GF180MCU-PDK; die Agentenoptimierung bewertete zwei Zellen. Es sind vom Anbieter gemeldete Entwicklungswerte für repräsentative Aufgaben.","en":"Before-and-after values reported in NVIDIA's 14 October 2026 Siemens EDA case study after refining agent skills and MCP tool descriptions. The benchmark used 229 standard cells and 36 PVT corners from the GF180MCU PDK; agent-tuning rollouts evaluated two cells. Results are vendor-reported development-benchmark figures for representative tasks.","es":"Valores anteriores y posteriores a la mejora de las habilidades del agente y las descripciones de herramientas MCP, comunicados en el caso de Siemens EDA publicado por NVIDIA el 14 de octubre de 2026. La prueba utilizó 229 celdas estándar y 36 esquinas PVT del PDK GF180MCU; las iteraciones de ajuste del agente evaluaron dos celdas. Son cifras de desarrollo comunicadas por el proveedor para tareas representativas.","fr":"Valeurs avant et après les améliorations des compétences de l’agent et des descriptions d’outils MCP, rapportées dans l’étude de cas Siemens EDA publiée par NVIDIA le 14 octobre 2026. Le benchmark utilisait 229 cellules standard et 36 coins PVT du PDK GF180MCU ; les itérations de réglage portaient sur deux cellules. Ces chiffres de développement sont ceux du fournisseur pour des tâches représentatives.","pt":"Valores antes e depois do aperfeiçoamento das competências do agente e das descrições das ferramentas MCP, divulgados no estudo de caso Siemens EDA da NVIDIA em 14 de outubro de 2026. O benchmark utilizou 229 células padrão e 36 cantos PVT do PDK GF180MCU; as iterações de ajuste do agente avaliaram duas células. São resultados de desenvolvimento divulgados pelo fornecedor para tarefas representativas.","tr":"NVIDIA'nın 14 Ekim 2026 tarihli Siemens EDA örnek olayında bildirilen, ajan becerileri ve MCP araç açıklamaları iyileştirilmeden önceki ve sonraki değerler. Kıyaslama GF180MCU PDK'den 229 standart hücreyi ve 36 PVT köşesini içeriyordu; ajan yinelemeleri iki hücre üzerinde değerlendirildi. Sonuçlar satıcı tarafından bildirilen, temsilî görevlerle sınırlı geliştirme kıyaslamasıdır."},"sourceUrls":["https://developer.nvidia.com/blog/optimizing-agent-reliability-and-tool-calling-efficiency-for-chip-design-using-nvidia-nemo-gym/"],"title":{"ar":"التغيرات المقاسة لوكيل Siemens EDA","de":"Gemessene Veränderungen beim Siemens-EDA-Agenten","en":"Measured changes for the Siemens EDA agent","es":"Cambios medidos en el agente de Siemens EDA","fr":"Évolutions mesurées pour l’agent Siemens EDA","pt":"Alterações medidas no agente Siemens EDA","tr":"Siemens EDA ajanı için ölçülen iyileşmeler"},"type":"metrics","updatedAt":"2026-10-15T18:50:44+00:00"}],"version":1},"data_downloads":[{"formats":{"csv":"https://newsverum.com/api/v1/articles/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026/data/siemens-eda-agent-evaluation?lang=fr&format=csv","json":"https://newsverum.com/api/v1/articles/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026/data/siemens-eda-agent-evaluation?lang=fr&format=json"},"id":"siemens-eda-agent-evaluation"}],"data_types":["metrics"],"editorial_illustration":{"alt":"Illustration conceptuelle de la technologie, des processeurs et des informations connectées.","asset_id":"newsverum-v1-technology","avif_variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.avif","width":480},{"height":480,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-720.avif","width":720},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.avif","width":960},{"height":800,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1200.avif","width":1200},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.avif","width":1536}],"credit":"Newsverum","documentary":false,"height":1024,"kind":"symbolic-editorial-illustration","language":"fr","scope":"topic","url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.webp","width":480},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.webp","width":960},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","width":1536}],"width":1536},"event_id":"event-ba230e4e-4307-5b5d-a84e-fc9a511265dc","event_period":"2026-10-14","event_period_basis":"event-or-reference-period","id":"nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026","image":null,"importance":0.6,"language":"fr","modified_at":"2026-10-15T19:00:45+00:00","publication_basis":"publication-record","published_at":"2026-10-15T19:00:45+00:00","reading_minutes":2,"region":"world","section_illustration":{"alt":"Illustration conceptuelle de la technologie, des processeurs et des informations connectées.","asset_id":"newsverum-v1-technology","avif_variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.avif","width":480},{"height":480,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-720.avif","width":720},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.avif","width":960},{"height":800,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1200.avif","width":1200},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.avif","width":1536}],"credit":"Newsverum","documentary":false,"height":1024,"kind":"symbolic-editorial-illustration","language":"fr","scope":"section","url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","variants":[{"height":320,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-480.webp","width":480},{"height":640,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-960.webp","width":960},{"height":1024,"url":"https://newsverum.com/assets/newsverum/illustrations/v1/technology-1536.webp","width":1536}],"width":1536},"slug":"nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026","sources":[{"checkedAt":"2026-10-15","name":"NVIDIA Technical Blog, Siemens EDA case study","publishedAt":"2026-10-14","url":"https://developer.nvidia.com/blog/optimizing-agent-reliability-and-tool-calling-efficiency-for-chip-design-using-nvidia-nemo-gym/"}],"summary":"Dans un benchmark de développement rapporté par NVIDIA, Siemens EDA a réduit les tours de l’agent de 155 à 53 et fait passer le score composite de vérification d’environ 55 % à 95 %.","text":"Dans un benchmark de développement rapporté par NVIDIA, Siemens EDA a réduit les tours de l’agent de 155 à 53 et fait passer le score composite de vérification d’environ 55 % à 95 %.\n\nNVIDIA indique que Siemens EDA a utilisé NeMo Gym pour tester les fichiers Liberty générés par son agent IA Fuse EDA pendant la caractérisation de cellules standard. Un vérificateur contrôlait la couverture des cellules et conditions d’exploitation, la validité des fichiers et le respect des tolérances des valeurs de temporisation et de puissance. L’équipe a étudié les échecs puis modifié par itérations les compétences de l’agent et les descriptions d’outils Model Context Protocol.\n\nDans le benchmark de développement rapporté par NVIDIA, le score composite de vérification est passé d’environ 55 % à 95 %, les tours de l’agent de 155 à 53 et les jetons de sortie d’environ 81 000 à 29 000. Le dispositif couvrait 229 cellules et 36 conditions du kit de conception de procédé ouvert GF180MCU, tandis que les essais de réglage de l’agent portaient sur deux cellules ; l’équipe a mené environ 25 itérations. Ce sont des résultats de NVIDIA et Siemens sur des tâches représentatives, pas une mesure générale de la productivité en conception de puces.","title":"NVIDIA décrit l’évaluation d’un agent de conception de puces avec NeMo Gym","topics":[],"translations":[{"api_url":"https://newsverum.com/api/v1/articles/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026?lang=tr","language":"tr","url":"https://newsverum.com/tr/news/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026?lang=en","language":"en","url":"https://newsverum.com/en/news/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026?lang=ar","language":"ar","url":"https://newsverum.com/ar/news/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026?lang=fr","language":"fr","url":"https://newsverum.com/fr/news/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026?lang=es","language":"es","url":"https://newsverum.com/es/news/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026?lang=de","language":"de","url":"https://newsverum.com/de/news/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026"},{"api_url":"https://newsverum.com/api/v1/articles/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026?lang=pt","language":"pt","url":"https://newsverum.com/pt/news/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026"}],"url":"https://newsverum.com/fr/news/nvidia-nemo-gym-improves-chip-design-agent-evaluation-october-2026"},"version":"1"}