Question Zen 6 Speculation Thread

[OneEng2]

No need to blame the entire readership for my ignorance

When I see RAS I read "Row Address Strobe" . I had to look up what it meant in the context of this discussion.

 

[coercitiv]

I would expect a fair number of garden variety layfolk frequent these parts nowadays. Not many well populated and well managed forums around anymore.

I object, I've been here for more than a decade already.

 

[igor_kavinski]

I object, I've been here for more than a decade already.

Now you are joking pretending to be layfolk.

 

[511]

waiting for people from 2000s to object

 

[Markfw]

waiting for people from 2000s to object

Like me ? Yes I object

 

[blackangus]

No need to blame the entire readership for my ignorance

Nice Job Igor!

 

[blackangus]

Wow, surprised the Anandtech readership is so ignorant when it comes to RAS features. This is standard fare on workstation/server level stuff. Intel and AMD even support it in their consumer memory controllers but it is (unfortunately) typically locked in retail units. IIRC memtest86 even supports it on the units that have it unlocked.

There are also DIMM interposers that allow you to inject errors not only in the data lines but also the control/checksum so you can verify that not just the data but also the commands are properly protected and test the memory controller's scrubbing capability.

I know tons of server side IT people that wouldnt know that level of detail.
Remember that your level of desire to know how things work is likely not generally shared, and alot of people never get to know this level of detail unless it breaks something and they have to learn!

 

[igor_kavinski]

Nice Job Igor!

Hey, my entire brand since birth has been "let me have a crack at it despite not knowing much and maybe it will work?" and it has served me well so far in life

 

[Mopetar]

Hey, my entire brand since birth has been "let me have a crack at it despite not knowing much and maybe it will work?" and it has served me well so far in life

Generally not a bad way to go through life, but as they say, "If at first you don't succeed, skydiving may not be for you."

 

[moinmoin]

Remember that your level of desire to know how things work is likely not generally shared

That lack of desire is a very bad thing.

alot of people never get to know this level of detail unless it breaks something and they have to learn

Do they though? My impression is only people with above desire would be interested in finding solutions and learn the details that way. Most people just give up at that point and move on to the next toy instead.

 

[igor_kavinski]

That lack of desire is a very bad thing.

It's not lack of desire as much as it is the difficulty in obtaining the required information. People with deep technical knowledge don't make it easy for normal people to understand their stuff. Like, I would only know about threads and processes if I had studied Andrew Tanenbaum's book on the subject. I actually did try but it was pretty boring and made it harder for me to go through it and I only read enough of it to pass exams.

I learned a great deal more from Anand Lal Shimpi and his team than any textbook. So while the lack of desire might be there in most people for delving into the deep details, it could be magnified by other more worrisome factors that are part of the life for most humans. Ideally (and how it used to work) was that people like Anand would work their butts off to scratch their curiosity itch and then write about it for the rest of us in a really easy to understand manner that would educate everyone thus leading to a more informed population.

Sadly, that tradition seems to be dying with more and more websites (even normal news ones) asking for donations. I also believe that Nvidia's stock price inflation is directly proportional to the general information dilution that has given rise to a much younger, dumber masses of asses. Telling or speaking the truth has always had a cost but it seems these days it's gotten so bad that extremely few want to bother.

 

[Schmide]

Andrew Tanenbaum - I found my original MINIX disks with manual rotting away in my garage. Legendary status.

 

[blackangus]

That lack of desire is a very bad thing.

I dont disagree here, but most people just want something to work and don't care HOW it works.

Do they though? My impression is only people with above desire would be interested in finding solutions and learn the details that way. Most people just give up at that point and move on to the next toy instead.

If its your job you can't give up, if its not your job then yes most people just give up.
In the context of RAS features and function its going to be someones job that make them learn that or the desire to know how things work.

 

[Doug S]

I know tons of server side IT people that wouldnt know that level of detail.
Remember that your level of desire to know how things work is likely not generally shared, and alot of people never get to know this level of detail unless it breaks something and they have to learn!

I would argue it is the server IT people's JOB to know this level of detail. If their employer is buying all prepackaged servers from Dell or whoever, along with any upgrades of DRAM etc. then it is Dell's job to be selling qualified configurations so I can excuse those admins.

If their employer buys whitebox servers or goes to random vendors for DRAM upgrades then it is their job to verify this stuff, you can't just assume it works. But I suppose a lot of them skip that or don't know they should be doing it.

That's hardly the worst oversight I used to see I suppose, but the people who "don't want to know that level of detail unless ... they have to learn" are to blame for the length of many outages. The absolute worst time to be learning something new is when everything is down and the C suite is breathing down your neck!

 

[fastandfurious6]

guys after zen 6 release and exponential advent of AI none of this will matter......


I think most of us got excited with hardware and OC at times when computers were slow.... even most expensive Extreme Edition™ pcs were not enough... no one could run Crysis ultra unless huge crazy SLI setup etc...

hardware was lagging behind software......

but now software gonna be lagging behind hardware

computers are so fast it doesnt matter anymore... and cutting edge advances will become so complex there wont be much a point to know deep dive hardware intrinsics unless u work at that level in a major hardware company i.e. amd nvidia qualcomm samsung intel etc

 

[511]

And people will make react native apps to cancel the hardware advancement made

 

[Fjodor2001]

The speed of software halves every 18 months.

 

[Josh128]

The speed of software halves every 18 months.

If you think about it, this has to be true, lol.

 

[igor_kavinski]

If you think about it, this has to be true, lol.

Yeah they do it on purpose just so you have to upgrade to keep up.

Win11 Start menu pops up with noticeable delay on older hardware than the Win10 Start menu.

 

[igor_kavinski]

no one could run Crysis ultra unless huge crazy SLI setup etc...

It didn't really matter because even on Core 2 Quad with a Geforce 9600 GT, I was able to enjoy it without much issue. The graphics were just so good that playing on lower settings didn't make the experience any worse.

The two games I remember that made me want to upgrade (but I couldn't) were F.E.A.R. and Riddick: Escape from Butcher Bay. And now these run at 100+ fps even on Ivy Bridge era Quadro GPU. I was surprised to see Half Life: Lost Coast running smoothly on Haswell mobile iGPU. That was a game that made many gamers sad with their existing hardware at the time.

 

[Doug S]

The speed of software halves every 18 months.

Fjodor's Law

Now we can all say we were present at the birth of a new computing law!

 

[Fjodor2001]

Fjodor's Law

Now we can all say we were present at the birth of a new computing law!

Actually it’s usually called Gates’s law (or Wirth’s law), see this.

Some spin-offs also mentioned on that page:
"What Intel giveth, Microsoft taketh away" or ”Andy giveth, Bill taketh away”

 

[Thibsie]

Fjodor's Law

Now we can all say we were present at the birth of a new computing law!

The speed of Microsoft software halves every 18 months.

maybe

 

[blackangus]

I would argue it is the server IT people's JOB to know this level of detail. If their employer is buying all prepackaged servers from Dell or whoever, along with any upgrades of DRAM etc. then it is Dell's job to be selling qualified configurations so I can excuse those admins.

If their employer buys whitebox servers or goes to random vendors for DRAM upgrades then it is their job to verify this stuff, you can't just assume it works. But I suppose a lot of them skip that or don't know they should be doing it.

That's hardly the worst oversight I used to see I suppose, but the people who "don't want to know that level of detail unless ... they have to learn" are to blame for the length of many outages. The absolute worst time to be learning something new is when everything is down and the C suite is breathing down your neck!

I wasnt saying what was right, I was saying what I actually see in most cases! =)

 

[LightningZ71]

The sad part is that, what makes computer systems feel more and more anemic as time goes on isn't that the OS itself is slowing down, it's all the constant security vulnerability mitigations that keep getting piled on. The computers that felt super quick on early Windows 10 at work a few years ago are just crawling these days. You crack open performance monitor and dig around and you see all sorts of auditing activities, antivirus scanning, memory randomization efforts, intentional cache dirtying, memory address scrubs, access profiling etc.

Systems don't really idle that often. There's always something in the background with something to do.