Hacker News Reader

Comment by bjoli | Hacker News Reader

Why no compilation benchmarks? Say, a Linux, chromium or GCC compile. All that cache must make it fantastic at compiling.

View on HN

Ever since moving to Win 1909, our compile benchmarks have been a bit off. I was at Supercomputing last week, literally got back Saturday to start writing the review, and I need to get some time to debug why it's not working as it should. I've got Qualcomm's Tech Summit next week and IEDM the week after that, so you'll have to wait a bit. Ian (the editor of the review)

Quick question: How are you handling Intel's security patches vis-a-vis benchmarks/consistency? In particular as there are like three(?) performance impacting fixes now that were released over a two or more year period.

For example in articles like this where you include older chips are you just listing the numbers from previous benchmarks, or are all running on the same patch level/OS?

I can't resist asking something that's been bugging me for a long time. The issue of insane travel schedules has openly plagued AnandTech for a long time now, both delaying reviews and severely impacting their quality when they do come out. Every tech website out there covers the big events in roughly the same way, very few do proper deep dive technical reviews like you've done in the past. Is it really the best use of your time to (partially) squander the comparative advantage you have in favor of rushing from one event to the next, reporting the same things everyone else does?

If you don't show up, no one knows you exist. He has to do everything the other guys do and more in order to stand out.

His CPU reviews in particular have at times been virtually unparalleled in the industry. I don't think name recognition is an issue. It's just painful to see that suffer for run of the mill industry news reporting.

Think of it as networking.

Ian,

Consider adding Linux as the underlying platform for some of your tests. Not only it's usually easier to script repeatable developer-oriented workloads, but your reviews will be more comprehensive as well. The Linux scheduler is quite different and more advanced (as illustrated by vastly different Gen 2 Threadripper performance on Windows vs Linux), and your articles have not been capturing this.

I can see why Linux is not yet the mainstream for gamers, but this OS should be more popular among the HEDT crowd.

Phoronix posted some compilation benchmarks: https://www.phoronix.com/scan.php?page=article&item=amd-linu...

And golem.de (link in german, see table at the end of the page): https://www.golem.de/news/threadripper-3970x-3960x-im-test-a...

The percentage increase in performance in those benchmarks is stunning. They're pushing hard on the multi-thread end and it's clearly working for a lot of modern applications that can take advantage of it.

Pair one of these Threadrippers with a pair of the fastest NVMe SSD's out there, with plenty of high performance DDR4 memory and you've got a near supercomputer from recent past in terms of performance.

We should also really update all tutorials out there to use things like `make -j12` and makefiles to link and compile using multiple threads.

make -j will do it, but having played around a bit via machines on packet.net, RAM is the limiting factor.

> If the -j option is given without an argument, make will not limit the number of jobs that can run simultaneously.

On large projects you'll run OOM quickly if you're compiling thousands of files at once. All the context switches also slow down compilation.

Really interesting and effective way to lay out the results of so many tests here.

https://openbenchmarking.org/embed.php?i=1911248-HU-THREADRI...

It's 2019; I just want to press F5 to refresh my Linux kernel. Is that too much to ask for?

I just want vscode to work, because it's awesome when it does... but I tend to have 5-10 editors open at a time and so far only sublime_text works w/ my workflow without crashing...though it sometimes needs rebooted when the plugin manager is having race conditions or other errors. Plus other electron apps. Current system is about 8-10 years old though, upgrading after christmas sometime to 3700x amd line and ddr4 3200 ram 32gb (upgrading from 8gb)...mobo should be quite upgradeable though to last another decade.

I just did this upgrade. Having so much compute at your disposal is magical. You won't regret it!

> mobo should be quite upgradeable though to last another decade

Wouldn't count on it. I can't remember the specifics, but we're coming up on the end of the vague time range that AMD said they'd support the AM4 socket. They talked about how difficult and expensive it was to get the interposer to connect the new chiplets to the old pins on the socket. I think Lisa Su also implied in an interview that AM4 is nearing it's end. In any case, DDR5 is coming in the next year or two which would necessitate a new socket, so I'd guess one more generation on AM4 at most.

I'm getting MSI MPG 570x motherboard. It's upgradeable to 128GB, 3700x is pretty performant. I don't game. I use linux + i3wm. I simply want my code editors/browser windows to stop crashing lol...

My goal is a decent build < 1000. I already have a case/power supply. Just getting CPU, Mobo, RAM basically. Keeping my r9 270x amd gfx card from 2013. Normal cooling. My RAM can go higher than 3200, but price range 3200 seems best bang for buck, I figure when it comes down I could upgrade to 128GB @ 4400 and eventually the cpu to whatever the highest end am4 socket allows, and a better graphics card and get a few more years out of my system.

I've never spent more than 800 on a system. I guess one man's 'wow this is cool' is another's 'no, that's lame' lol.

Ah. I assumed that "upgradeable motherboard" referred to being able to upgrade the CPU. My mistake.

AMD has already confirmed that Zen 3 (Ryzen 4000) will be supported on AM4, so there's at least one generation left on that platform.

Last time I checked on this, the promise was a lot less clear than everyone is making it up to be.

Elsewhere, Zen 3 has been rumored to be a full rewrite, i.e. a new microarch from the ground up. That doesn't jive well with maintaining socket compatibility with AM4.

The words I've seen from AMD on this matter seem fairly clear that they intend to keep Zen 3 on AM4. Most likely they'll wait until Zen 4 on a 5nm process to switch to a new socket(in 2021).

It's still a 'bit' upgradeable. But I think RAM/Gfx will get me further over the next 7-8 years in terms of speed/etc. It's a decent build comparable to maybe a $2000+ consumer desktop for $699. For my use case It'll fly compared to my system now which bogs down when I have 126 browser tabs open.

Why would you not upgrade an 8-10 year old mobo? They aren't that expensive.

Gamers Nexus did some GCC benchmarks with the regular Ryzens and indeed they were absolutely trouncing intel, and indeed it was due to the big l3 cache.

The Threadrippers are the same cpu with just more of them so I'm sure they do well also.

On Linux: https://www.youtube.com/watch?v=oRgmq9f6VBY

https://www.level1techs.com/article/threadripper-ragnarok

Note that every core has access to "only" 16MB of cache (well, counting L2 you could say it's 16.5MB). This will change with Zen3 ("32MB+").

Reply on news.ycombinator.com