QXC

How exactly are you all running your firefox compile benchmark? I wanted to run myself for a basic comparison but my build times are way faster for some reason. I did make sure not to use artifact mode but I can't find any info about how the test is ran by LTT / any other reviewers.

The point being that semiconductor companies want their high performance parts to be on the smallest node possible by default. TSMC makes the smallest node, so that's where they go. Making chips faster and more efficient is what the company does by default, people on the internet have no influence on that.

The supply chain could be a bit better for sure, but I don't think that's entirely on AMD. Microcenter wasn't saying ANYTHING about if they'd even have any in stock until a few hours before they opened today. The only way for stuff being sold like this to be "fair" is to do it by lottery. Otherwise it's either an arms race of who has the best online tools to snag pre order listings as soon a they exist or just being fortunate enough to live near a brick and mortar store that has them in stock. Admittedly, all issues that could be solved with greater supply. But then, how late do you delay a product to wait for good dies to pile up? They still don't do high clocks, but they ARE on a much newer (smaller) node. But they certainly didn't do that because people on the internet were complaining.

That quote regarding the stock in the UK was immediately put into question on reddit. No conclusive evidence either way but I'd say it's more speculation that fact in any case. 3950X launch hasn't had huge numbers of chips available, but I'd hardly call that a "failure". As long as demand outstrips supply, it's a good tech product launch. It does sting a bit if you miss out on the first batch, admittedly. If we're going to start being upset at companies for not having stockpiles of product on hand for every launch, let's just be upset evenly regardless of who does it. And at least a little glad that this wasn't a pure paper launch.

So, I don't entirely agree with TPU's conclusion based on their own benchmarks. So maybe I'm missing something. The basic CPU tests still come in behind the G5600 on average, with some exceptions. IGP performance was ahead of intel as one might expect (Although both systems were running high end ram for sake of even comparison) but also WAY behind the 3200G. Unfortunately they don't have a G5400 to compare against, and the price premium of the G5600 makes it look awful in everything price wise. But in this price range even a few dollars can make a massive difference in value and/or performance. At this particular moment in time, one can get a Ryzen 1500X (4C8T) on Amazon for $69, or if you are blessed enough to be within range of a microcenter you can pick up a 2200G for $59. I feel like the performance boost in those instances is worth stretching the budget, or waiting for sales. All of that being said, while the 3000G is a competitive chip, I don't feel like it warrants a glaring recommendation. It should be about even with a G5400. Mildly faster if you OC, much faster for IGP, but also stomped on by chips mildly more expensive than it. Even the text in the review kinda comes to this conclusion: But as for beating the Intel offering easily, I'd argue it's more of a toss up outside of IGP. Worth noting that all of the better value options in similar price range are AMD, the G5400 is the ONLY thing intel has on the low end that remotely makes sense. TPU claimed verifying with AMD, and the awful frequency scaling would indicate Zen 1. Totally worth the clarification though, since it would make the naming way less obnoxious if it was zen+.

It's unlocked, but it's zen 1 so the overclocking potential is kinda damp squib. Yeah there's SOME headroom but it requires an OC to kinda compete against a similarly priced intel chip.

This is two generations back instead of one.

Article: https://www.techpowerup.com/review/amd-athlon-3000g-with-vega-3-graphics/ I haven't seen a ton on this outside of TPU, they don't touch on it much but I find the naming scheme a bit misleading. I guess it's commonplace since stuff like this happens to varying (Nvidia GTX 250) degrees (AMD R7 300) with other (Intel 109X0) hardware, but I really would have preferred they call this the 2000G. Or have just not made the APUs start at 2 instead of 1. In other news the chip isn't amazing, I guess.

OP: "We interviewed Lisa Su and she said that the 4000 series APUs (Zen 2) would be on AM4 and released early next year!" Replies: "Wow the Ryzen 4000 CPUs (Zen 3) will be on AM4 early next year!"

When did we decide that websites and browsers should be applications and operating systems rather than documents and document viewers?

This leak is pretty much the ideal situation for what Intel could do. Price cuts plus a 10 core part, but oops you need a new motherboard again. The TDP numbers are interesting but ultimately irrelevant, and I think everyone would be surprised if there was a magic generational power efficiency improvement with the same arch on the same process. I don't even think Intel would be brash enough to just lower their TDP numbers when it's obvious that the processors aren't really any different. But hey, maybe they'll drop the actual announcement around the time all of those red hot AMD chips drop later this month.

Less than a month later stock is trickling back in and prices are mostly back to normal. Newegg still trying to skim a bit off the top and amazon selling out first, as usual.

you rite edited

https://nordvpn.com/blog/official-response-datacenter-breach/ The official response from Nord with more details about the incident. This appears to refute the opening commentary from techcrunch. Or at least attempt to.

Maybe. But are the chips actually going to be for sale that day? Or is it going to be when the new TR chips actually go on sale? Or a different date? And even those days are rumors still. I want something official. "Go out to microcenter on this day and deposit money."

LGA 775 would like to have a word with you. But that's on a technicality as back compat between P4 era 775 and Core era 775 was spotty. Still waiting for a hard release date on the 3950X

Has it been confirmed that the 3700X and 3800X hit similar frequencies when you do this? From what I've read/recall, the 3800X is a much better binned die and generally clocked better than the 3700X, but I don't know for sure if that was with similar power limits.

You don't remember the first dual core CPUs in the mid 00s, do you? Or just generally how CPU performance went until Intel hit the clock speed wall and suddenly we had to care about how much power we threw at our CPU.

Thanks AMD! I don't think they can price this competitively until AMD runs out of 2000 series (And 1000 series!) stock. The 9100F (4C4T) is going for $90 and that's just not a good buy when $20 more will get you a 2600. Or a 2700 for well under $200. When those are gone, the sub $200 market will be more manageable. I think they'd need to price this around where the 9100F is now and put an overclockable version around $150 to stand a chance. But by the time intel brings these to market, AMD will probably be close to launching the Zen 2 based APUs, and those might have a huge advantage if intel keeps slicing their IGP off of desktop chips.

The topic was allowed on the forum until it devolved into a political discussion, which they touched on during the show. This one might too, so I'm gonna put my "In before the lock" here.

Wan show should only read superchats and FP chat Oh and wan show VODs should go on FP please please please don't make me use youtube for twitch for that

Some might argue that Skylake/Kaby Lake chips are still in production, albeit with different model numbers...

My understanding is that large portions of the current Intel GPU team were poached from people that helped make the Kaby Lake-G project. Don't know how much truth there is to that, but there are certainly some former AMD graphics engineers at Intel now. New Intel GPUs are going to have teething problems. Just like every other new thing does. Will be interesting to see if they turn out to be not garbage, given how many times they've tried this before in one form or another.

While amusing, it's very logical to conclude that their deletion system/workflow simply wasn't designed to handle a huge amount of users at once, which appears to have happened with the recent news. So, if anything it's a good sign that users are actually willing to take a stand against their actions. Deleting an account might not seem like much, but it shows the strongest possible sign internally that the userbase is not happy with what's going on. Plus I'm sure "active accounts" or "total accounts" is some sort of metric that they use internally or with investors, and a huge dip in that is a sign that their actions are actually having a financial impact on the company, not just a bunch of people moaning online between overwatch rounds.

Source: https://www.anandtech.com/show/14970/intel-releases-advisory-update-x299-for-cascadex-and-lose-kabyx-support So lots of easy jokes about how Kaby Lake-X was a horrible idea and a great example of how Intel's marketing team isn't always the brightest. But, does this mean that you can't use a Kaby Lake-X chip to update the bios on say, the new board you bought that didn't get an update for Cascade Lake-X out of the box since it had been sitting around on the shelf for a while? Or can you update it but as soon as it updates you can't use the old chip anymore? That's kinda the only valid use of one of those chips now, kinda like the <$50 Athlons on AM4. Ultimately not actually big news. As fun as it is to go "Hey look, Intel motherboards are having the same problem as AMD motherboards!", Kaby Lake-X was "very niche" at best and I doubt they moved many of them. For that one person out there that's been sitting on a full X299 platform and a quad core waiting for something like the new price cut specials, good luck with your bios update. EDIT: For clarification, this is not an Intel mandate as far as I know, it's just an issue of space on the BIOS/UEFI chip. So just like with the AM4 platform, it's hit or miss depending on how big that chip is and who your MB vendor is.