Speaking of computer history:
https://blogs.msdn.microsoft.com/directx/2019/01/07/wall-of-gpu-history/
https://files.facepunch.com/forum/upload/108618/81c2e66a-fa74-42ab-9c2d-b7f3c924ff3c/image.png
Radeon VII
Not what I was expecting but better than what nvidia wants for a 2080 :/
8c/16t zen 2 CPU beat a i9-9900k whilst using less power in cinebench, magic
7nm vs 14nm, it damn well better. Now I just hope that the singlethreaded performance compares favorably as well and that the prices won't be as disappointing as VII.
Do I get to say "I told you so"? Because I totally told you so. This is pretty much exactly what I expected of Zen 2 - a solid improvement over Zen+, better IPC, probably better clocks, but not nearly as big a leap as Bulldozer -> Zen, and no massive core count inflation while cutting prices.
Radeon VII is nifty, I don't need a GPU upgrade yet (my 1080Ti is maxing out every game I own at 1440p60) but AMD's competing at the high-but-not-top-end again, which can only be good for the consumer.
I hate RGB but I somehow ended up with an RGB case with RGB light strips with an RGB CPU cooler complete with RBG Fans and RBG Mouse and Keyboard
I've become what I have despised
They are gimmicky and not worth the price when you don't want them.
But once you get them, they are really cool and versatile to have since you can just change your color scheme any time. RGBs are cool when you're buying components that are top-tier, otherwise it's a waste of money.
I think its mostly because all of the top tier hardware and gear is just RBG by default, the only thing I went out of my way to get was the RGB strips, and they're actually the most practical RBG gear I have when you need to see into your case to make a quick adjustment
Beating it in Cinebench
Yawn.
All I can say is atleast they're making true progress in CPUs. Their GPUs might be a joke, but atleast it's a legitimate competition in CPU space. They can make money now and not be dying every damn quarter.
Raja's aftershocks.
The one distinct advantage it has is 16GB VRAM. Which do you think is more likely: RT/tensor becoming useful in the amounts present on RTX2000-series cards, or textures swelling to over 8GB? I'd put more money on the latter than the former.
hard agree on this. what intel is delivering year after year is fucking profoundly disappointing.
Monitor calibration tool; good or bad investment?
I noticed that the Spyder 5 Express/Pro weren't as expensive as I thought. If these things last a long time and are accurate, then roughly 150$ doesn't sound too bad if it means I'll be able to accurately calibrate most of my screens. Or is it not as simple as that?
unless you have a serious professional need for a monitor calibration tool, I personally would spend $150 on something else. do you go through new screens constantly? do you have an abundance of screens that are actually high-quality enough to justify using a professional calibration tool? a calibration tool is only going to slightly polish a turd if the monitor itself can't cut the mustard. unless you are a graphic designer with a wide array of monitors that you MUST work on and ensure consistency throughout, I can think of other things I would buy first
Yeaah, a color calibrator is overkill for my needs when thinking about it like that. I don't switch out my monitors until years have passed and the only use case I really have for a accurate monitor is some photography I do rarely and my monitors are already good enough for that. Thanks for the input, I'll look out for better monitors versus a color calibrator in the future when that's due for an upgrade
I have uncovered content.
I'm coding an integration with Fedex. They have some sample code, including PHP which is what I'm using (seriously, is there anything better for mostly front-facing sites?). Unfortunately, that sample code is absolutely evil.
Behold, their attempt at central configuration, and in particular, how it capitalizes "Return":
function getProperty($var){
if($var == 'key') Return 'xxxxxxxx';
if($var == 'password') Return 'xxxxxxxxxxx';
if($var == 'shipaccount') Return 'xxxxx';
// snip 150 lines of this pattern
}
I don't even know what this is trying to do, but I know it's doing it the wrong way:
function setEndpoint($var){
if($var == 'changeEndpoint') Return true;
if($var == 'endpoint') Return 'xxx';
}
This line hiding near the top of the "fedex-common.php5" library, at top scope, so it just shits a debug message into your response:
echo 'PHP_Common initialized ';
And some comments that outright lie:
define('SHIP_CODLABEL', 'CODexpressreturnlabel.pdf'); // PNG label file. Change to file-extension .pdf for creating a PDF label (e.g. CODexpressreturnlabel.pdf)
Overall, I can't fault it too much. Like this isn't the worst PHP code I've ever seen, not by a long shot. But man is it needlessly hard to follow. And why the hell is the return keyword capitalized?
Uh yes, tonnes. What are you making?
Probably won't stay here very long, but I made the top 100 on the single GPU Port Royal RTX benchmark:
3DMark Port Royal Hall Of Fame
https://files.facepunch.com/forum/upload/106970/40594fce-2503-49cf-b9ea-a9356e79f3d8/image.png
I guess it would be classified as a business process system? Kind of like an ERP, except without the whole inner-platform-gone-mad thing, and aimed at "let the peons do their work" rather than "show the boss a bunch of graphs".
It's PHP mostly because I'm technically a consultant on this job, and the "lead" dev is exclusively a PHP guy, but I honestly don't know what I would have rather done it in. It's done as a web thing because clients have access to parts of it, and so it's about 80% input forms, remaining 20% mostly PDF generation. Other than ASP I don't know of any languages really well-suited to this sort of thing, where you're mostly just procedurally spitting out HTML.
And that's three-for-three with my certs! A+, Net+, and MCP, all first try.
Also the first person at the facility to pass the Win10 version of the MCP and did so without the proper training, because the Win7 MCP exam was retired midway through the MCP course without Microsoft notifying us. I was the first one who even got this far to find that out and now I'm the first to pass, all while having to play catch-up to new standards on short notice.
Any backend language...?
Setup Spring Boot or Node with templates I'd imagine, shame the lead dev is a PHPer.
I haven't played with Node much, but it seemed to me like it was built much more for the "everything is an AJAX request from a web application" school of programming than the "load HTML straight from the server" school, which is just not what this project needs. Maybe I just looked badly, or it's changed since, but the templating engines for Node were pretty bad. Or at least, they didn't work at all the way I wanted them to work, and I really don't like fighting my language or tools.
Pretty sure any templating engine that PHP supports Node also supports.
PHP basically is a templating engine, I literally do not understand why people would layer another one on top of it, other than programmers' innate need to cause problems for themselves by reinventing wheels.
Any text not in <?php ?> tags is sent as output. You need to echo a variable, <?=$foo;?> does that as cleanly as any templating engine. You need any logic, conditionals or loops or whatever? You've got all the language constructs of PHP.
Wheels like readability, code quality, performance?
Man, fuck subversion so goddamn hard.
How is a templating engine any faster or more readable than just embedding PHP code?
What's any better about:
<div id="username">{{username}}</div>
versus:
<div id="username"><?=$username;?></div>
At that point you're mixing business logic with templating, so you've lost your decoupling.
When your template is nothing but a template, you can swap out a theme for another without needing to make changes to your business logic.
Loops and conditions can still be part of presentation logic, not business logic.
For instance, here's some details-trimmed code from our generic table view, used on a bunch of things that were basic enough to not need a hand-coded view. Given $items as an array* of model objects, and $fields as an array of the fields we want to display, this gives a perfectly usable table:
<table>
<tr>
<? foreach($fields as $field): ?>
<th><?=$items->GetFirst()->GetColumns()[$field]->label;?></th>
<? endforeach; ?>
</tr>
<? foreach($items as $item): ?>
<tr class="<?=$item->GetListRowClass();?>">
<? foreach ($fields as $field): ?>
<?=$item->DisplayListElement($field); ?>
<? endforeach; ?>
</tr>
<? endforeach; ?>
</table>
I don't really see how that would be done better with a templating system. Isn't the more template-y way to do that some junk with fragments, just to move the loops into the controller code?
* Actually a ModelGroup object, which is a thin wrapper I wrote around an array to set the array key as the model ID, do some extra validation, and add some utilities like GetFirst() and GetLast().
Sorry, you need to Log In to post a reply to this thread.
