本帖最后由 BFG9K 于 2023-1-23 01:26 编辑
发布会是这张图啊，那里有平均15%的提升了？
这挑的四个游戏里，战锤3和DOTA2本来就是缓存敏感型游戏，没30%的提升就不正常好吧？
至于CS:GO, 58X3D对比58X是负提升，所以78X3D对比58X3D的提升其实是源于提频和ZEN4 IPC，和缓存没半毛钱关系
还有R6S,这游戏58X3D对比58X只有在1080P下才有15%的提升，缓存边际效应明显
所以78X3D拉开和58X3D在R6S上的提升，还是有频率和架构的因素在里面的
然后这是AMD 78X3D官网的footnote
Game testing as of 5 December, 2022, by AMD Performance Labs using the following hardware: AMD Socket AM5 Reference Motherboard with AMD Ryzen™ 9 7950X, Ryzen™ 9 7950X3D and G.Skill DDR5-6000C30 (F5-6000J3038F16GX2-TZ5N) with AMD EXPO™; ALL SYSTEMS configured with NXZT Kraken X63, open air test bench, GeForce 4090, Windows® 11, PCIe® Resizable Base Address Register (“ReBAR”) ON, Virtualization-Based Security (VBS) OFF. All games tested at 1920x1080 with HIGH in-game preset and the chronologically newest graphics industry API available within the game’s rendering engine (e.g. Vulkan® over OpenGL™, DirectX® 12 over DirectX® 11). Desktop configurations will vary, yielding different results. Games tested: CS:GO, League of Legends, Assassin's Creed: Valhalla, Grand Theft Auto V, Hitman 3 Dubai GPU, Wolfenstein Youngblood (LabX), Hitman 3 Dubai CPU, Cyberpunk 2077, Borderlands 3, DOTA 2, Middle Earth: Shadow of War, F1 2021, Far Cry 6, Final Fantasy XIV, Warhammer: Dawn of War III, Shadow of the Tomb Raider, Total War: Three Kingdoms Battle, Deus Ex: Mankind Divided, Watchdogs: Legion, Rifbreaker CPU, Red Dead Redemption 2. RPL-035
搭配4090在1080P下测的，但实际上1080P网游的真实需求用4070/3080已经封顶，用4090的BUFF下提升还需要D2这样的title才能拉到30%，我是觉得对比58X3D提升不大。
毕竟搭配4090测，58X3D对比58X的差距也能暴力拉升到MAX 49.4%, 30%+的title TPU测了都有9个，78X3D官方cherry pick了才1个30%+，还只是一个dota2, 不觉得很拉跨吗？ |
310112100042806
发表于 2023-1-22 23:50
如果访存完全不是瓶颈，那么频率和IPC的提升对应的性能提升计算起来就非常容易了，不会出现访存延迟或者带宽瓶颈导致高频IPC显著下降的问题。Zen4再怎么牙膏也有8%的IPC提升，再叠加11.1%的频率提升，保底20%了吧。
