Missed PC release deadline by 5 years. Good luck with that.

Willing to bet money that the co-op multiplayer will be as broken and unplayable as in Megaton edition, or even worse. They probably designed the new episode without co-op in mind, too, because, well, everyone forgot how to do it.

The episodic release model has been a disaster. Between your episodes, people forget the freakin' controls and game mechanics, not to mention forget about the game. Square Enix, you can go suck it.

WoW emulation was created by the community. I am not sure how these servers on that site are ran, because last time I checked, the only feature-complete emulation of WoW is "WoW classic" 1.12, and the biggest server of that is Kronos 2, which features thousands of people, and is free.

JohnBirshire wrote on Aug 31, 2016, 10:55:

shiho wrote on Aug 31, 2016, 05:06:
Watching videos of what's become of WoW, all I can say is Thank Jeebus for the community painstakingly preserving WoW classic via emulation.

Pfft, community, hah.

There are companies, not communities, making millions of dollars a year hosting WoW servers. www.warmane.com They have WotLK, Cataclysm, and MoP servers. More people online than you'll find on any WoW server as well. For example, one of their WotLK servers has 12,000 people online right now. Many WoW servers have 1-2,000 if they're lucky.

Watching videos of what's become of WoW, all I can say is Thank Jeebus for the community painstakingly preserving WoW classic via emulation.

When Duke3D Megaton edition, aka "the definitive playable version" came out, without promised multiplayer, and then kept delaying and pushing it back, I predicted that they are stalling and will never have functional multiplayer.

The developer got pissed at me and banned me from Steam forum. Then, eons later, my cousin and I tried the alleged multiplayer. It was a buggy, laggy mess - and we lived in the same city.

Now they pulled that game from Steam store entirely, to erase the shame. Guess someone's trying again?

Meanwhile, I recall, there was one way to actually play Duke3D in co-op, and it was to install a very specific old stable version of eDuke32, but I don't recall the version number.

SlimRam wrote on Aug 26, 2016, 05:27:

Jerykk wrote on Aug 26, 2016, 04:44:
DSOGaming has already confirmed that this is just yet another re-release of Duke 3D, with some minor enhancements and new content.

Ah so it's not a remaster but more like a re-release? Balls...nevermind I guess

Sepharo wrote on Aug 25, 2016, 21:15:
There's a whole series of screen capped posts of that one trolling the other on his facebook posts.

http://imgur.com/gallery/2Vvy9

Make sure to load the remaining 6 images too.

A low IQ individual being mocked by a medium IQ individual. Not sure who to feel sadder for.

@descender: As a former decent Tribes: Ascend player, I had a noticeable advantage with ping of 30 over ping of 40. If you were less stubborn than you are clueless, you'd start doing your own research. Tragically, the two qualities in you are equipotent. Good luck with that.

_________________

Mordecai Walfish wrote on Aug 16, 2016, 15:57:
stuff

1) IPv6 should be disabled when not used. In my anecdotal experience, doing so improved Realtek LAN stability. I followed Internet advice.

2) QoS should probably be enabled, though I still haven't figured out exactly how that shit works. This is what I theorize at the moment:

* You need QoS to minimize instances when you copy files on the LAN and they choke movie streaming on the LAN.

* The only way it MAY work (MAY), is if you set "QOS: Do not use NLA" to "1" in TCP Optimizer, and set the number above it to 20, which appears to be percentage of reserved bandwidth in order to prevent stream saturation.

* Priority&VLAN setting in your adapter should at least have Priority enabled, because this is linked to QoS packet tagging (?!)

All in all, this is how I set things based on the information I have.

3) RSS is one of those "too complicated to work" features. Applying multithreaded processing to single-thread processes (aka incoming network traffic), is heavily going to depend on programming skill of whoever implemented it. And it looks like half of it is coded by the driver team of Taiwanese child slaves.

On my LAN disabling RSS fixed streaming movie hiccups. I was breaking my brain trying to figure out what was happening, with 20x more bandwidth being available on the LAN than was needed for streaming a 480p video file, and during one scene it just kept going to 15fps. Disabled RSS, went away. Enabled it, 15fps again.

Many people's computers have C-states enabled, as well as SpeedStep.

Plug into that the complexity of multi-core processing... juggling various shut-down parts of the CPU... I just... from programming perspective this just doesn't sound like something they can get right.

The other statement you quoted, basically amounts to this: "your CPU is too slow to deal with low buffers, but RSS makes the processing faster, so you don't lose packets". I am not sure that's how things actually work.

After the RSS part, it's Windows traffic management that has to deal with flow control in software. Whether RSS is enabled or not. Right?

Or, you can try enabling Flow Control in the adapter itself - there's conflicting advice about it, as it may interfere with Windows' traffic management, but on the other hand, it could save CPU cycles?

I dunno, this stuff sounds iffy, especially considering that online gaming traffic doesn't put much demand on data processing in the first place. And if you're gaming, you probably made the effort to shut down your Bittorrent. I'd rather make sure the packets are processed in proper order, in a streamlined fashion.

And if you do run Bittorrent or anything TCP-based, really, you're not going to lose much by dropping packets. The control mechanisms work pretty well.

4) I don't know about buffers. It's a weird thing. I am not sure whether buffers are actually processed in linear fashion or not. I noticed that decreasing buffers on the Intel card to the minimum (80), alleviated SOME of the packet aggregation created by the shitty power ethernet adapter, and gained some responsiveness in Tribes: Ascend.

However, if the power-ethernet wasn't in the way doing its garbage, it wouldn't have made much of a difference. I also don't know exactly how much data a "buffer" contains, and whether it is universal among manufacturers. Realtek 8139 had that setting in Kilobytes, so I understood it, but now I don't.

In general I set receive buffers to 256 and send buffers to 128 to achieve necessary balance between responsiveness and not overloading the system with flow control handling.

Right now I am using an Intel PCI card with buffers of 80/80, aforementioned QoS-related stuff enabled, everything disabled except for TCP and UDP checksum offload.

Because it's Intel, and their brand of Taiwanese slave programmers are known to be less retarded.

I also disabled "Windows Heuristics" for handling TCP adaptivity window in TCP Optimizer. It's like a manager on top of a manager, Office Space-style.

And yeah, ECN is highly questionable because it can mess with your actual Internet traffic. The world is not ready.

Unfortunately a lot of this material is scattered around the Internet in bits and pieces, so that's my current "reality" of it. It may change.

This comment was edited on Aug 17, 2016, 23:08.

Mordecai Walfish wrote on Aug 16, 2016, 14:24:

Pigeon wrote on Aug 16, 2016, 13:37:
Though I wouldn't be surprised if its also so they can more easily hide crap that consumers would otherwise not want. Sounds like it would eliminate the ability to not install update XYZ.

Hit the nail on the head. They have been trying more and more inventive ways of suggesting users of Windows 7 and 8 to upgrade to 10, and take part in their "telemetry services" through windows updates. I keep a list of the KB numbers to avoid and always prevent those from installing. With the update structure they seem to be moving to here, that likely won't be possible. I'm going to have to make a slipstream installer for windows 7 with windows update disabled, if this turns out like I expect it to. Fucking MS.

Technically it should be still doable. Someone needs to keep a blacklist of malicious updates, and add them to a growing .bat file which you run after every Microsoft update pack.

descender wrote on Aug 16, 2016, 00:13:

No. Disabling Nagle's and TCP Checksum Offload were popular fixes because of WoW.

So, yes? Didn't I just say that?

No. You implied that ALL of the settings I mention only apply to games like WoW. You also claimed that they only affect TCP protocol, which is incorrect.

Increased CPU access is merely one of the possible side-effects of one of these setting changes.

I don't want to harp on this topic all week. Believe what you want.

The side-effects is what happens when those features are enabled.

You really should've stopped this a while back, when it became evident you are poorly informed. Maybe you can stop now?

If you use old drivers, old NIC's and old OS's then maybe some of these settings will help. Sure, they all technically "decrease latency", it just doesn't translate to anything tangible in the real world.

Given how Interrupt Moderation is packet aggregation - meaning more information is buffered before it gets processed - it does affect gaming response time. Glueing together several ticks worth of gameplay data, that otherwise had a chance to be responded to, has its consequences, regardless of whether your in-game latency checker shows it or not.

So do broken checksum calculations, which result in retransmissions.

Your 10ms sugar pills are delicious! :)

You can in fact gain 10ms depending on how fucked up your network adapter is. And 10ms is huge.

Edit: It would appear that you've been posting this same information on other message boards for over 3 years. 3 years ago they were much more important changes to make.

Actually, the faster computers get, the less reason it is to rely on manufacturer's hacked-together code squeezed into "lowest-cost" hardware, and all the more reason to make the hardware do only the bare minimum, and offload all those extra "features" onto the "spec" traffic processing code which is part of Windows itself.

descender wrote on Aug 15, 2016, 15:52:
Very few of these settings will make any difference at all for twitch online gaming (though there is definitely some merit to disabling nagle's and RSS).

Everything shino has said here is technically correct about the TCP protocol. TCP sucks and it definitely requires all of these tweaks to lower latency. The problem here is that no games that you actually need or want low latency for actually use TCP as the protocol. They all use UDP for those connections (BF, CS:GO, OW, etc) because it doesn't require packet acknowledgments. Developers are very aware of the TCP limitations and that's why they don't use it.

Games that you will see the most marked improvement on from the changes he suggested (WoW, Guild Wars) use TCP for a reason... mainly that your world position isn't as important and they can operate within the latencies of TCP without issues.

These things became "popular fixes" mainly because of WoW, and because of the old NIC's/drivers that were lying about their onboard feature sets. They really don't apply to a majority of twitch gaming today.

No. Disabling Nagle's and TCP Checksum Offload were popular fixes because of WoW.

All adapters have both TCP and UDP checksum offload, however. And when it's broken, which it often is, it will affect UDP traffic.

Interrupt Moderation, probably the most lag-causing feature, is not limited to TCP traffic, either.

So, your claim that my suggestions only affect TCP games, is wrong.

You can try the tweaks without harm, but there are side-effects to disabling nearly every one of these settings.

No, the only side-effect is potential increase in CPU time used by the adapter. But since onboard garbage made by Realtek, already offloads to CPU - as you stated - the difference would be acceptable for most people. The biggest impact would be created by disabling interrupt moderation, I'm guessing, but it is also the biggest culprit, as it is a packet aggregator.

If you use an Intel card, you may want to leave TCP/UDP checksumming enabled, as it in fact can save CPU time if you're in a really high-traffic environment.

I should have been more clear though. I don't run TCPOptimizer monthly to keep the gaming tweaks intact, I do that to keep the TCP stack optimized for my 70Mbit connection (driver updates will undo lots of settings and bandwidth is largely irrelevant for gaming).

I create my own "optimized" presets in TCP Optimizer, and then load them on various machines.

descender wrote on Aug 14, 2016, 09:16:
Blanket changing of those NIC settings was more important in the past with older drivers, OS's and slower CPU's. Now NIC's basically do no work on their own and offload everything to the CPU anyway regardless of your settings and these should not all be disabled.

I do recommend running SG TCP Optimizer on "Optimal settings" for all adapters though, that will fix any misconfigurations you may have acquired and will clean up your connection more than any of the other changes will. You should only disable the extra offloading features if you are having excessive driver or latency issues.

These latency changes are miniscule at best and would be unnoticeable without direct measurement. If you get a drastic improvement here it was because one of your other TCP/IP settings was probably wrong like your MTU/RWIN/etc. I try to run the TCPOptimizer like once a month just to make sure nothing in windows changes the settings (which it likes to do with driver updates). You don't have to change any individual setting though, the "Optimal" set works just fine.

Correcting the settings can make a dramatic difference in twitch online FPS. Interrupt moderation is a biggie. Others are often broken or follow half-agreed on specs.

It's not going to be felt in WoW, but there were cases were checksum offload broke WoW as well.

The motherboard NICs may not offload to hardware, but they do offload nonetheless. They offload to their own checksum calculation code, made by their shitty programmers.

This is why there was a scandal about Realtek driver corrupting data. When you disabled the offloading features, the bug went away. You know why? Because it went back to letting Microsoft code handle it, aka, "the spec".

This is the safest mode in most circumstances.

As for the revered "TCP Optimizer"... where do I start. Its "Optimal" settings enable "TCP Chimney Offload", which is Large Send Offload, which is broken as a concept, and is pretty much regarded as universally bad.

They also enable "Receive Side Scaling", another wonderful feature responsible for tons of problems. If you're streaming video off your NAS server and it stutters in some scenes, this feature could be the culprit.

If you're a gamer, both of these MUST be disabled.

The amazing "Optimal" setting also enables Direct Cache Access on Windows 7/Vista, a feature that was scrapped on Windows 8 and up, because it wasn't properly supported by vendors.

"Optimal" according to whom, exactly?

Mordecai Walfish wrote on Aug 14, 2016, 14:56:
My NIC is a killer 2200 built into my motherboard. I think they do alot of that buffer removal stuff already, and I use TCP optimizer on optimal settings already.

Any tips I should know for the 2200?

Thanks

I'm not familiar with that one. However, see above reply in this post about TCP optimizer.

Dev wrote on Aug 14, 2016, 03:11:

shiho wrote on Aug 14, 2016, 01:29:
Unless you have an Intel card, you should disable pretty much every single feature in your network card

And if you do have intel? Can you leave stuff enabled?

Intel usually has reliable checksum offload. However, large-send offload (also known as Chimney Offload), is dysfunctional as a concept, and there's also been rumors of Intel breaking it especially bad.

If in doubt, disable pretty much everything. Flow control is a kind of a crapshoot, sometimes hardware works better with it enabled, other times it will disrupt Windows' built-in traffic management.

If you do enable it, it should be enabled on everything on the network. Intel actually seems to disable it by default.

_________________________

BOTTOM LINE: you may want to experiment, and disable all features in TCP Optimizer and your network driver which have the following words in them:

Offload
Moderation
Flow
Scaling
RSS
Green
Energy-saving
DCA
TCPA

I think that's about it.

This comment was edited on Aug 15, 2016, 22:53.

Should've kept the 20hz. Now the game will, most likely, yield the tick count when someone's running a download which is competing for bandwidth. After all, most people don't know how to set up QoS.

And of course, via this "adaptivity" it discriminates against people with less bandwidth.

The shot-behind-walls effect and general feeling of everything being behind a step was always there. When I found out the root cause, it made sense that was the case, and so I rated the matches I played 1/3 to reflect that I think they needed to move beyond 56k modem settings with their tick rate.

Actually that is not the root cause. 20hz has been standard in FPS forever. Some even used 10hz. This would not manifest itself in lagging a step behind. It manifests itself in, if player movement is circular, it may instead resemble an octagon. The more updates per second, the smoother the curve gets.

What does manifest itself in the way you described, is:

a) Generally shitty netcode, without pre-step lag compensation and other modern features.

b) Packet buffering on the level of your network adapter, switch, or router. For example, most consumer ethernet cards, including those built into your motherboard, aggregate packets. The feature is called "Interrupt moderation", and should be disabled by every gamer.

Unless you have an Intel card, you should disable pretty much every single feature in your network card, AND in Windows as well, via TCP optimizer. Everything with the word "offload" in it (this covers a lot of broken features). Flow control (yes, flow control is broken on Realtek cards). Green Ethernet.

If you use a ethernet-via-power adapter, they all aggregate packets, and there's almost nothing you can do about it. You can set the send/receive buffers to 64 or 80, depending on whether you have Realtek or Intel, and it will reduce it, but that's about it.

This comment was edited on Aug 14, 2016, 01:37.

"This qazaj is no good", sighed John. "Look". He pointed at the station wall, as Guaskforo squeaked nervously behind him, and pulled what, for all intents and purposes, seemed like a trigger. A tiny light danced on the reenforced steel, making an "X" mark, and then fizzled out.

"See? Doesn't work."

The eyegoggled trader in front of him, started speaking in agitated gibberish.

"Dbod sot roein. Fas fas Bgarr!"

"I got one word out of that", said John. "Arrival".

"Well, I'm sorry", he sighed. "I can't wait for you to figure your shit out anymore, Guask. Hey, where you off to?", he quizzed, following the pudgy alien awkwardly plodding down the corridor. "You have a vraz in here for me that I don't know about?"

"BGARR!"

"What? English, motherf-"

He didn't see the wall behind him shimmer, and the darkness push through. John suddenly lost his footing, being pulled down the corridor by an incredible gravitational force. He grabbed and held on to nearby panel for dear life, watching Guaskforo's ugly face through the transparent sphere which closed around him. Lucky son of a bitch.

John screamed. He screamed twice.

Guaskforo turned away, as quashed organics colored the forcefield crimson red, just for a moment.

He pressed his cortex to the receptivity matrix, and the Bgarr around him shook, as it was catapulted from the station. Guaskforo looked up, watching the station be consumed by Dbod.

These Earthlings. They never learn the language. It's their own damn fault.

This comment was edited on Aug 11, 2016, 16:04.

Not a wholly authentic Soviet experience.

HorrorScope wrote on Aug 10, 2016, 12:22:
Does ComboFix work with Win10? I'll check and run it if so. I get your point, well noted and if these weren't mostly entertainment systems yes I would run a tighter ship. But I do measure the systems crticalness and the results I do see and adjust over time. I'm not saying I'm right, I'm not telling others not to do whatever makes them sleep at night and all that.

I will say is you are very confident in what you mention and seem to shit on most AV systems in general it seems, but I have to think then almost all people using AV's are running AV's that you possibly deem as no good, so look at all of that false hope. I want real hope, if it is there for window dressing, I really have no need for it. If it can really do the job, then I wants.

As I said before, Combofix does not support Win8 and higher. Here you are stuck with booting from something like Hiren BootCD in its own mini-Windows7 (if it works), and then hoping it recognizes your Win10 partitions and network card so you can update one of the scanners on it (like MBAM), and then scan your system "from the outside". That should negate a part of the rootkit stealthing efforts, however cumbersome and inconvenient it is.

_______

AVs are not "no good". I never said that. They are a limited part of the overall solution. Using Windows Defender is almost like having no antivirus at all, so it is BETTER to have a decent one with realtime filesystem protection and web traffic monitoring enabled.

There's actually one promising AV called I think Webroot(?), which actually tracks changes made by unknown executables, and reverses them if they're proven to be harmful.

That's some real innovation. I just don't know whether it can track ALL the changes they make, or whether viruses can work around that too.

The best game Carmack has made in the past 19 years, has been Wolfenstein RPG, specifically for dumbphones. The phones which weren't necessarily flipphones, but not smartphones either. No hardware acceleration, VGA screen resolution.

I almost passed the entire game, got stuck on the final boss. But I had TONS of fun with it.

Why? Carmack clearly had fun (FOR ONCE) going back to his roots and working in a confined hardware environment.

Not only did it allow him to focus on tight, engaging gameplay (and sense of humor), but he managed to squeeze in subtle effects like software alpha-blending, on phones without any 3D acceleration.

Finally, unlike its fancy iPhone version, this game played better, because I still had the luxury of pressing the real, hardware directional buttons to navigate the world and perform actions. People have forgotten... but it makes a HUGE difference to get tactile feedback on your actions.

Wolfenstein RPG > RAGE

HorrorScope wrote on Aug 9, 2016, 16:11:
What if I scan the system periodically and it comes up clean? I don't run real time, but I do scan every so often, haven't found anything in years.

What I've been trying to get through to people for YEARS, and which doesn't seem to sink in, is that post-factum "scanning" doesn't do shit anymore.

Either the attack is stopped before it hits the vulnerable software, via a TCP/IP scanner, or it gets stopped during the execution of the injected or just user- or program- ran code.

If it doesn't, the virus becomes a rootkit. From that point on, it is undetectable by conventional antivirus solutions, which lag 10 years behind industry needs in that department.

They mask themselves using filesystem tricks, append to system files, replace system files, etc. Some detect when infected files are being scanned, and un-infect them, then re-infect them. A lot of those tricks were used in the DOS era, and later forgotten.

ComboFix has been the one program that actually "roots" them out. It is basically an automated hacker toolkit designed to destroy harmful hacker toolkits.

If you run Win7 or XP, I suggest you have System Restore enabled, and then run it.

Windows Defender is the worst of the worst.

Called it. More excitement to follow.

@HorroScope: That's the other extreme. Running without a decent antivirus (and MS shit doesn't count as such), is reckless.

It IS possible to get infected, like for instance with the recent hack of a download site which embedded a virus into the installers.

Or, through a hole of some shitty browser addon like Flash.

Or, say, you bought a Ubisoft game, realized what a hassle their "service" is, and decided to pirate it instead, because pirates get a better version of it, with the service stripped out. You run a questinonable crack made by some kid with morals looser than his mom's vagina.

Worse yet, in the past 10 years the nature of Windows viruses changed drastically. Practically every virus these days is a rootkit.

Once it is executed, you will have no traces of its activity in the system, such as CPU time expenditure. Maybe excessive bandwidth spending - MAYBE NOT.

Conventional antiviruses are as great at dealing with rootkits as an ED-209 was at going down the stairs.

There are specialized utilities like Combofix, which do a much better job. Unfortunately Combofix doesn't support Win8 and up.

I would suggest you run ComboFix (from BleepingComputer.com) for shits and giggles. Let it create the System Restore point first, of course. See if it digs something up in its deletion report, because it very well might.