英伟达宣布，推出基于Ampere架构的新款工作站显卡，分别为RTX A1000和RTX A400，两者都采用了极简的设计风格。
这是英伟达时隔两年半以后，再次推出Ampere架构专业显卡，此前最后一款产品是RTX A2000 12GB。从官方提供的渲染图就能看到，RTX A1000和RTX A400都采用了简洁的设计，厚度只有单槽，而且都不需要任何外接供电，TDP都控制在了50W，发热量应该都不大。
与基于Ada Lovelace架构GPU的产品不同，RTX A1000和RTX A400缺乏AV1编码支持，不过可以对包括H.264和H.265等其他视频格式进行编解码。英伟达尚未确定RTX A1000和RTX A400的定价，前者将很快上市，后者要等到下个月。
var tableid = settableid();
document.write(”);
Table().init({
title: “NVIDIA Ampere 桌面工作站显卡”,
header: [“型号”, “RTX A6000”, “RTX A5000”, “RTX A4500”, “RTX A4000”, “RTX A2000(12GB)”, “RTX A1000”, “RTX A400”],
width: ‘92%’,
columnWidth: ‘MAX’, //表格单元宽度，”MAX”最大内容显示，””AVG”平均，”AUTO”自动，指定[15,16,20]
id: tableid,
data: [
[“GPU”, “GA102”, “GA102”, “GA102”, “GA104”, “GA106”, “GA106”, “GA107”],
[“核心面积”, “628mm²”, “628mm²”, “628mm²”, “392.5 mm²”, “276 mm²”, “276 mm²”, “200 mm²”],
[“工艺制程 “, “三星8nm”, “三星8nm”, “三星8nm”, “三星8nm”, “三星8nm”, “三星8nm”, “三星8nm”],
[“CUDA Cores”, “10752”, “8192”, “7168”, “6144”, “3328”, “2304”, “768”],
[“Tensor Cores”, “336”, “256”, “224”, “192”, “104”, “72”, “24”],
[“加速频率”, “1455 MHz”, “1697 MHz”, “1653 MHz”, “1536 MHz”, “1200 MHz”, “1463 MHz”, “1758 MHz”],
[“单精度浮点(FP32)”, “31.2 TFLOPs”, “27.8 TFLOPs”, “23.7 TFLOPs”, “19.2 TFLOPs”, “8 TFLOPs”, “6.74 TFLOPs”, “2.7 TFLOPs”],
[“显存容量”, “48GB GDDR6”, “24GB GDDR6”, “20GB GDDR6”, “16GB GDDR6”, “6/12GB GDDR6”, “8GB GDDR6”, “4GB GDDR6”],
[“显存位宽”, “384-bit”, “384-bit”, “320-bit”, “256-bit”, “192-bit”, “128-bit”, “64-bit”],
[“显存带宽”, “768 GB/s”, “768 GB/s”, “640 GB/s”, “448 GB/s”, “288 GB/s”, “192 GB/s”, “96 GB/s”],
[“NVLINK VRAM”, “96GB GDDR6”, “48GB GDDR6”, “40GB GDDR6”, “N/A”, “N/A”, “N/A”, “N/A”],
[“TDP”, “300W”, “230W”, “200W”, “140W”, “70W”, “50W”, “50W”]
]
});