Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

CPU Usage Differences After Applying Meltdown Patch at Epic Games

epicgames.com

Post reply on HN
Print view View original

281–290 of 296 posts

Re: CPU Usage Differences After Applying Meltdown Patch at Epic Games

#281

Post by pixl97 »

cookiecaper wrote: post #197
drawnwren wrote: post #76

Earlier quoted context omitted.

This is a horrible approach to security. If you only secure against attacks you expect, you're gonna have a bad time.

Some things are just not applicable. There's no point adding another padlock to the outside wall of Fort Knox. If untrusted users don't send instructions to the same physical CPU that you're using, Meltdown/Spectre are not relevant. I also think that we need more information before we really see the long-term performance impact here. First, the patches will likely be optimized over time. Second, I am not sure that PT…

In the past if you had a RCE that only got user level access and that user was very locked down there was only a limited amount of damage they could have done. Now with meltdown, every RCE is a full information loss at system level exploit.

Re: CPU Usage Differences After Applying Meltdown Patch at Epic Games

#282

Post by pixl97 »

jedisct1 wrote: post #66

The Meltdown patch also introduces a serious performance hit for DNS servers and resolvers. Before giving figures, I want to run the same tests on a more recent CPU, but my current benchmarks are not great to say the least.

Is this a Program -> CPU interaction that is slowing things down, or is this a CPU -> network interaction.

I'm wondering if there are classes of network drivers that are having a much larger effect in performance. Network cards these days can do many things to improve performance like TCP/UDP offloading and because of that their drivers are very complex and I'm going to assume that there well be Meltdown fallout because of this.

Re: CPU Usage Differences After Applying Meltdown Patch at Epic Games

#283

Post by dxhdr »

Thaxll wrote: post #155
contrarian_ wrote: post #21

Pretty much what I predicted here: https://news.ycombinator.com/item?id=16054674 > Sounds like servers handling lots of small UDP packets would be hit pretty hard.

Gameservers don't handle that many packets because they're limited in the number of player they host. A game a 60hz with 64 players will only receive 3840 packets/sec.

And then how many game servers can you run on a machine? Hopefully a lot more than just one.

Re: CPU Usage Differences After Applying Meltdown Patch at Epic Games

#284

Post by snuxoll »

zzzcpan wrote: post #88
_wmd wrote: post #47

Earlier quoted context omitted.

There's potential for a little rearchitecting to help, at least in the case of UDP: NAME sendmmsg - send multiple messages on a socket SYNOPSIS #define _GNU_SOURCE /* See feature_test_macros(7) */ #include int sendmmsg(int sockfd, struct mmsghdr *msgvec, unsigned int vlen, unsigned int flags);

If they are using socket API for UDP, performance is not critical for them. Otherwise porting UDP servers to DPDK/netmap is not rocket science and gets you like an order of magnitude better performance.

A shooter running at 30Hz (high side for free developer sponsored servers) with even 100 players is only going to be processing at most 10-15K packets per second, and that’s assuming 5 packets per tick per player. Server update rates are usually lower than the internal physics and game logic tick rate as well, so it’s doubtful to even be that high.

These aren’t stats of the gameplay servers, these are the backend servers that handle matchmaking, player stats, inventories and progression.

Re: CPU Usage Differences After Applying Meltdown Patch at Epic Games

#285

Post by snuxoll »

serf wrote: post #185
sillysaurus3 wrote: post #168

Earlier quoted context omitted.

Hopefully not. Lawsuits are a fine way to stifle innovation. Imagine how hard it would be to push through any idea at Intel. No one saw this coming. Things happen. It's impossible to predict every contingency. They acted in good faith. Also, heh, users are funny: "all other games i have work fin by the way so there must be a problem whit fortnite."

>They acted in good faith. That's new for a company, and it's not indicated by their PR spin right now. What I think happened : a company produced a product with a problem. Probably not out of malice, but ignorance. One of two things happened after, which can kill the 'good faith' argument ; the problem was found internally and hushed, or the problem was found externally and minimized to reduce financial burden arisi…

Intel got a report about this vulnerability from Google in July. Intel's CEO decided to sell stock in November, scheduling the sale in October. Intel also decided to pull the Coffee Lake desktop launch in from early 2018 all the way back to the start of Q4 2017 to try stop the momentum AMD was building with Ryzen, while knowing this vulnerability was present - they're still planning on launching Cascade Lake the first half of this year and god knows if meltdown will be fixed in it or not.

Right now, I think the problem started out of ignorance - but they have abused Google's policy of responsible disclosure to hide the flaw as long as they could and take advantage of their market position while the unknowing public kept buying their products. Now they are pulling the four D's of propaganda in their PR statements all while we are seeing huge performance deltas in graphs from Epic and more?

This is straight up deceptive, I'm glad I switched back to AMD with my new gaming rig and I already have plans in the works to purchase multiple EPYC servers with our datacenter move starting next month.

Re: CPU Usage Differences After Applying Meltdown Patch at Epic Games

#286

Post by snuxoll »

dijit wrote: post #250
mtgx wrote: post #130

Looks like it may be time for Epic Games to consider switching to AMD Ryzen/EPYC.

I have been looking at AMD/Epyc CPUs recently. (Because making NUMA aware C++ code is hard, and AMD Epyc is a single socket on a server with 4 very closely knit NUMA zones so non-NUMA code will run better on that vs Intel) But unfortunately there's no comoddity server from HP/Dell available yet. But I hear one is on the way on the Dell side.

HP has paper launched the DL385 G10 with EPYC, I'm not sure if it's fully available through sales channels or not. Supermicro also has had EPYC systems available for a while - though that won't do you any good at all if you need a big name OEM for "nobody got fired for buying Cisco" reasons.

Re: CPU Usage Differences After Applying Meltdown Patch at Epic Games

#287

Post by urbanxs »

Abishek_Muthian wrote: post #60

If it's helpful, Our Node.js, MongoDB, Python servers all with significant network traffic didn't have any measurable impact after KPTI patches on Amazon Linux on T2.medium(burst), M4.large, T2.large(burst) respectively. Our impact is lesser than the figures suggested by redhat's advisory - https://access.redhat.com/articles/3307751

That’s most likely because neither node, mongo nor python actually use the optimisation features of cpus. Says a lot about their performance quality.

Re: CPU Usage Differences After Applying Meltdown Patch at Epic Games

#288

Post by therealdrag0 »

burnte wrote: post #274
yborg wrote: post #195

Earlier quoted context omitted.

"Significant network traffic" tells me almost nothing, that's a totally subjective description and you don't describe your workload at all. A back-end MMO game infrastructure server I have some understanding of in terms of the kind of load it is handling.

It tells us his workload is more network bound than computationally intensive.

Exactly, in my company's SaaS monitoring, I excluded db/external calls, and I saw a 1-2ms increase (on-top of 15ms), so yeah that's a 10% increase, but with external calls make the overall transaction 50+ms, so 1-2ms is trivial.

Similarly, the CPU on our Azure cloud box holding a half dozen dockerized Java apps I can't really see went up at all. Hovering around 10% CPU usage before and after patch.

Re: CPU Usage Differences After Applying Meltdown Patch at Epic Games

#289

Post by merb »

pixl97 wrote: post #280
merb wrote: post #271

Earlier quoted context omitted.

Well a lot of processes run as Root anyway. Also if possible i run as few processes as possible

>Well a lot of processes run as Root anyway. Then you didn't listen to DJB like 15 years ago. Moreso, your processes shouldn't be running under the same username. Principle of least user authorization people.

please just install a linux distribution and run ps aux. a ton of stuff just runs as root. either because there is no other way or because it's a bad default.

Re: CPU Usage Differences After Applying Meltdown Patch at Epic Games

#290

Post by msbarnett »

merb wrote: post #289
pixl97 wrote: post #280

Earlier quoted context omitted.

>Well a lot of processes run as Root anyway. Then you didn't listen to DJB like 15 years ago. Moreso, your processes shouldn't be running under the same username. Principle of least user authorization people.

please just install a linux distribution and run ps aux. a ton of stuff just runs as root. either because there is no other way or because it's a bad default.

Yeah it’s your job as a Unix dev or admin to not stick with crummy but convenient defaults, kid. Nothing exposed to the web or touching anything exposed to the web should be running as root. Each thing should be running under its own, very limited, user and group. If something gets hacked, you want it to be as least privileged as possible.

We know how Unix works. We’re telling you because you clearly are out of your depth. If you don’t understand that you have no business maintaining anything on the internet.

Post reply on HN