新的英特尔处理器 KPTI 错误。浮点计算会出现哪种减速？
New Intel processors KPTI bug. Which slowdown to expect for floating point computation?
一些媒体报道说在 Intel 处理器中发现了一个 new hardware bug，允许用户模式进程访问内核模式内存：
It is understood the bug is present in modern Intel processors
produced in the past decade. It allows normal user programs – from
database applications to JavaScript in web browsers – to discern to
some extent the layout or contents of protected kernel memory areas.
The effects [of fixes] are still being benchmarked, however we're
looking at a ballpark figure of five to 30 per cent slow down,
depending on the task and the processor model.
错误修复后，我预计多核浮点计算会出现哪种减速？
据我了解，只有内核模式和用户模式之间的切换性能受到影响。例如，处理大量 I/O 是一种经常出现的工作负载，但 CPU 密集型流程不应受到太大影响。
引用一篇分析 Linux KPTI 补丁性能的文章：
Most workloads that we have run show single-digit regressions. 5% is a good round number for what is typical. The worst we have seen is a roughly 30% regression on a loopback networking test that did a ton of syscalls and context switches.
...
So PostgreSQL SELECT command is about ~20% slower with KPTI workaround, and I/Os in general seem to be impacted negatively according to Phoronix benchmarks especially with fast storage, but not gaming performance, Linux kernel compilation, H.264 encoding, etc…
因此，如果您的 FP 计算主要依赖于内存中的数据移动而不是 I/O，它们应该基本不受影响。
一些媒体报道说在 Intel 处理器中发现了一个 new hardware bug，允许用户模式进程访问内核模式内存：
It is understood the bug is present in modern Intel processors produced in the past decade. It allows normal user programs – from database applications to JavaScript in web browsers – to discern to some extent the layout or contents of protected kernel memory areas.
The effects [of fixes] are still being benchmarked, however we're looking at a ballpark figure of five to 30 per cent slow down, depending on the task and the processor model.
错误修复后，我预计多核浮点计算会出现哪种减速？
据我了解，只有内核模式和用户模式之间的切换性能受到影响。例如，处理大量 I/O 是一种经常出现的工作负载，但 CPU 密集型流程不应受到太大影响。
引用一篇分析 Linux KPTI 补丁性能的文章：
Most workloads that we have run show single-digit regressions. 5% is a good round number for what is typical. The worst we have seen is a roughly 30% regression on a loopback networking test that did a ton of syscalls and context switches.
...
So PostgreSQL SELECT command is about ~20% slower with KPTI workaround, and I/Os in general seem to be impacted negatively according to Phoronix benchmarks especially with fast storage, but not gaming performance, Linux kernel compilation, H.264 encoding, etc…
因此，如果您的 FP 计算主要依赖于内存中的数据移动而不是 I/O，它们应该基本不受影响。