Minecraft Logic Circuits

You can make just about anything in Minecraft, and that includes the computer that you’re using to run Minecraft. Of course, your computer has billions of individual transistors, so that level of complexity isn’t quite feasible. People have managed to build fully-functional computers in Minecraft, but they require months (if not years) of effort and are at best comparable to their real-life equivalents from the 1970s. The smallest building blocks, on the other hand, are fairly easily doable.

Let’s start with what computers are actually doing on a lower level. Basically everything a computer does can be boiled down to either arithmetic or a load/store operation (in essence, reading/writing memory). When you hit your keyboard, an electrical signal tells your computer that you’ve done so, and it mathematically determines what this input means and how it should respond.

A fundamental component of a CPU is the arithmetic logic unit (ALU), and one of the most important components of that is a circuit capable of performing addition and subtraction. This circuit on its own is relatively simple, as it’s purely made of combinatorial logic. However, each block within the circuit can only operate on one single binary bit, so it’s necessary to string X number of them together (where X is the number of bits we need to be able to operate on).

The easiest way to do this is to determine at each individual bit whether or not there will be a carry-over to the next bit, then send that signal in as one of the next bit’s inputs. This is known as a ripple-carry adder. However, it can be quite slow. If you have a carry-over that needs to be pushed from the very first (least significant) bit all the way over to the last bit, it has to traverse through each bit’s addition circuitry before it gets where it needs to go.

One of the most popular alternatives is to split the circuit into sections, each with its own logic that determines the carry for each individual bit within as well as the section as a whole, the latter of which gets sent out to the next section. Since the logic determining these carry signals is relatively fast, the circuit can compute the sums of all bits at practically the same time, instead of waiting for the carry signal to ripple over from the first bit to the last. This implementation is known as a carry-lookahead adder.

Shown below is a 16-bit carry-lookahead adder displaying 23451 + 30575 = 54026 in binary. Note the distant output display, top center of the image. The total volume is over 72000 blocks.



When Drugs Stole my World (And How I Got it Back)

Quick disclaimer that this isn’t the main writer on TGN; I hope to share more insights in the future.

It’s been a long time, almost a year and a half. I forget a lot of things that happened, a lot about how I felt at the time, a lot of the wisdom I was giving to others at the outpatient. But I still have a story to tell. To show that what I did is possible; to show that with enough struggle, hard work, and trial and error, sobriety is achievable.

Without going into too much detail about how I got there, like most I started with pills and never thought I’d end up using heroin. They’re safer, right? You know what’s in them and how much, and doctors give them to people on prescription. (At least back then, this was somewhat true; I’ve heard most “oxys” going around nowadays have random amounts of fentanyl.)

This is not to mention that at the same time, I had serious problems with cocaine, and to a lesser extent Xanax. The combination produced a level of sheer pleasure that’s impossible to describe. But all of these habits needed to be tended to and paid for, consistently and without delay. That’s always the biggest problem isn’t it?

Unlike many others, I fully understood what kind of fire I was playing with — I had prior experience with all sorts of drugs, and had known both addicts and recovered addicts who told me their stories, but it never got quite so out of control. This time was different. I had just tried to kill myself and figured that going off into the “land of nod” might be an easier, more fruitful alternative.

But of course, pills are expensive, and they’re meant to be taken in small amounts for pain management. Me, I would do a month’s prescription in as little as two days. And, obviously, the dependency set in soon enough, and I needed bigger and bigger doses, more and more frequently. So once I had bought every last pill the dealers in my neighborhood could find, and all the other sources had me waiting for the next month’s prescription, it was a no-brainer what would come next: heroin. I needed something, anything.

Once you try heroin, in my opinion at least, there’s no going back to pills. Why would you spend ten times as much money for basically the same thing? Heroin is cheap and it’s available everywhere. I’d be scrambling all around the city looking for oxys, but dope was just a five minute walk. And if that dope man wasn’t around, the next one was only five minutes from there.

I never got as bad into it as some others. I was limited by what I could afford, so a decent bundle would have to last two days. And, as I mentioned before, every two days I was also going through an 8-ball of coke and 16-32mg of Xanax. I never really got into stealing to support the habit, with the exception of ripping people off on drug deals and disappearing with the cash. I did spend every dime I could find on it, though, and went deep into debt.

By the time I overdosed, I was sure I would put an end to it. I weened myself off a bit, going down to a bag or two a day, then got a ton of Xanax and hunkered down for the withdrawal. It was awful, but I hadn’t been in the game for too long, so it wasn’t as bad as it was when I quit for real. But this wasn’t lasting.

Instead of seeing the risk and avoiding opiates, I was emboldened by this whole situation: I felt as if I’d demonstrated that I can handle opiates, that I was able to cut it out when I decided it was time. So when I was offered oxys, I’d buy them. I had no intention of staying sober. I just wasn’t actively seeking dope.

Then one day, I got offered dope. It was pretty funny, actually. I ran into my man walking down the block, he said “yo, I got that food,” and that was that. Quickly I went back to the same old habit — bundle and 8-ball every two days or so, though with less Xanax involved. The debt I had paid up started to rack up again, and at this point all my bridges started to burn.

Eventually, it got to the point where the drugs weren’t working anymore. The whole point was to stop me from killing myself, but I had fucked up so bad that I realized I now actually wanted to kill myself more as a result of what the drugs did to me. On top of that, I was isolated, and felt everyone looked down on me. I was doing poorly at work and constantly in debt, doing crackhead shit and asking people for money to pay for the necessities that came after dope on my list of priorities.

Most people at this point just treated me like shit. People were always trying to mess with me, trick me, rip me off. Always being subtlety disrespectful and condescending. My family was constantly upset and treated me with great suspicion. A couple close friends begged me, crying, telling me to stop, talking about how much they worried about me, including one of the authors from TGN (before TGN existed). This had an effect, but it also made me sad, and sadness led to suicidal thoughts, and suicidal thoughts lead to escapism, and escapism lead to more dope.

Eventually, though, I realized I had to cut it out. I couldn’t afford my habits anymore and my family had learned all about them. I struggled for a couple weeks, arguing with myself and trying to use less without plunging into the hell of dope sickness. But at this point, it didn’t feel like much of a choice — I had to change. I was thousands of dollars in debt and on the verge of suicide again.

So, I started by cutting out the coke. After doing coke every waking hour of the day for weeks, if not months without a break, the crash can be pretty fucking brutal. In the past I had even experienced very serious signs of cardiovascular issues during these crashes. But I still had my dope, so it was actually manageable.

Then the dope came next. The promise of the sickness was terrifying. After a day, it was obvious that I couldn’t handle it. So I borrowed some money and got some kratom, knowing it would help. I started taking absolutely massive doses, but they took the withdrawal away about 90%. It was a lifesaver. But it also became another habit, albeit not a destructive or fiendish one, which I’m still ever-so-slowly tapering myself off of. In some ways, I attribute my being alive today to kratom. At the same time, part of me wishes I had been able to stop using it right after I was over the withdrawal, instead of using it as a long-term maintenance therapy. But that seemed like it would lead to me killing myself, and it might have, so I’ll take whatever works. I got clean, and I’m not dependent on conventional maintenance treatments like Suboxone or methadone (actual opioids which are far stronger and harder to quit than kratom).

That’s not to say there weren’t setbacks. A day or two after I stopped using dope, I lost my job as a result of the poor work I had been doing the last few months I was on it. I still had to pay my living expenses and pay off the colossal debt, so on top of that I had to turn to my parents for financial assistance. At this point I was completely destroyed and felt like I couldn’t get any lower. All pride, ego, narcissism, even basic dignity to an extent, destroyed and gone with the wind, never to recover.

A week later I finally got off the Xanax. That was more of an intermittent thing, though, and I hadn’t done too bad with the last round of it, so a quick self-taper was all I needed. I had been through far worse Xanax withdrawal before, so this wasn’t a big deal. I don’t even remember what it felt like coming off of them that time (well, in fact, I don’t remember much about my Xanax usage at all) but it was nothing compared to the dope sickness. Nothing could ever compare to that. (That’s not to say Xanax isn’t fucking nasty!)

Almost a month into sobriety, I was still experiencing physical withdrawal symptoms. Lethargy, low appetite, muscle pain. I couldn’t walk even 10 minutes down the street without simultaneously feeling like my leg muscles were tearing and like I was going to faint. The cravings for both coke and dope were absurd, causing muscle spasms and making me writhe in pain and scream at the top of my lungs for no apparent reason. On top of that, mentally, I hadn’t really gotten anywhere. The addict mind state is hard to break out of. This all combined with encouragement from friends and family led me into an outpatient.

Outpatient was a useful experience. Find a good place where you’re not going to find many junkies who are just in there because of a court order, and where the counselors are decent and actually care to understand you. Pay attention to what they teach and engage the others there who have some sober time and can give you good insight. Don’t be afraid to speak up and share. It’s my belief that this is what truly leads to recovery — you can stop doing drugs without it, but a good support program will get you out of the mental state of addiction, and that’s the most important thing. That’s what will prevent you from relapsing and enable you to move on with your life.

It took about a year of sobriety for my life to actually get better. And frankly, if you’d ask me most days, it still fucking sucks. But if I think about how much it sucked two years ago, I shudder. It feels like it’s behind me now, and I no longer have the same frame of reference. At this point, I couldn’t see myself ever relapsing, so long as I stay vigilant. Staying vigilant is key; you just have to keep yourself far removed from drugs and drug users and be aware of the fact that addiction is a lifelong disease that, in the wrong circumstances, you can fall back into again.

I used to use all those drugs to stop thinking about suicide. Now I’d rather just think about suicide than go through all the bullshit the drugs put me through. Most of the time, honestly, I would rather just kill myself if I was going to relapse, ironic since it used to be the other way around.

So, what’s the point at the end of this long-winded, rambling story? That if you or anyone you know is struggling with addiction, sobriety is absolutely within reach. All it takes is a radical shift in perspective. An understanding of the reasons why you use drugs, which then leads to the understanding that drugs actually make all of these problems worse. Most (some would say all) of us have to learn this the hard way. But if this story can be of any sort of assistance to even one person, that’s enough of God’s blessing for today.

Intel’s Misleading Thermal Design Power (TDP): Don’t Trust Your Processor’s Wattage Rating

Image credit: ewfewfev.blogspot.com

Intel has come under scrutiny lately for the power consumption and heat output of its CPUs, specifically the 9th Generation Core lineup consisting of Coffee Lake and Skylake-X parts. This much is understandable, as a revitalized AMD has forced Intel to increase core counts and clock their chips aggressively. However, the rating Intel gives its chips that’s supposed to inform users how much heat it outputs — the Thermal Design Power (TDP) — has not risen substantially in turn. Let’s just take a look at a cross-section of processors:

CPU Intel Core i7-7700K Intel Core i7-8700K Intel Core i9-9900K Intel Core i7-6950X Intel Core i9-7980XE Intel Core i9-9980XE
Node 14nm+ 14nm++ 14nm++ 14nm 14nm+ 14nm++
uArch Kaby Lake (Skylake) Coffee Lake (Skylake) Coffee Lake (Skylake) Broadwell Skylake-X Skylake-X
Cores / Threads
4 / 8 6 / 12 8 / 16 10 / 20 18 / 36 18 / 36
All-Core Turbo
4.4GHz 4.3GHz 4.7GHz 3.4GHz 3.4GHz 3.8GHz
TDP 91W 95W 95W 140W 165W 165W

The Core i9-9900K, despite being practically double the CPU and running at a higher clock speed on the same manufacturing node, somehow is only rated for an extra 4W of heat. This heat output rating is closely correlated, but not entirely equivalent to, the chip’s power consumption, as the vast majority of power drawn by the chip is dissipated as heat.

The secret to the formula is the CPU’s base clock, the lower advertised speed. In the footnotes, Intel’s TDP rating is technically only valid for the base clock. This is generally considered as a bottom floor for the CPU’s frequency under normal load scenarios. When idling, the chips run at a far lower frequency, and when loaded as long as there’s no thermal throttling (or AVX-512 instructions) they should run at a higher frequency. The i7-7700K has a base clock of 4.2GHz, compared to 3.6GHz for the 9900K. The 7700K only goes 200MHz above the speed for which it’s rated 91W, whereas the 9900K climbs a massive 1.1GHz.

Another element is the fact that TDP is generally an imprecise metric. Historically, since the initial generation of Core processors, Intel’s TDP has always been very liberal. The TDP rating was an absolute worst-case scenario; your chip was likely to consume dozens of watts less. The 7700K tends to consume up to 10W less than its TDP would indicate despite running at 4.4GHz on all cores instead of 4.2GHz.

Now, as you might have guessed, reviewers have generally found the Core i9-9900K to consume somewhere within the range of 150W and 180W of power under load with the default BIOS configuration. It’s possible to limit the CPU to its TDP in the BIOS, as is the case by default with many pre-built systems, but this results in lower sustained clock speeds and noticeably worse performance than reviews would indicate.

Why is an accurate TDP rating important? For one, it helps a user decide how beefy a power supply they need to run the chip. Secondly, it’s supposed to inform users — and OEMs like Dell and HP — how beefy a cooler is required to keep the chip at acceptable temperatures. In fact, Intel’s own decisions about which cooler to include in the box are based on TDP (the 9900K and 9980XE don’t come with coolers, hint hint). The Core i7-8700 has a TDP of just 65W and comes with a cooler rated for 73W for good measure. However, it actually has the same all-core turbo as the i7-8700K, despite a 500MHz lower base clock. If you try to run it with the stock cooler in a consumer motherboard, it will immediately overheat and throttle under load, as Tom’s Hardware has demonstrated.

AMD’s TDP for their Ryzen parts, on the other hand, is by all indications very accurate — which they ought to be commended for, even though they can’t reach the extremely high frequencies that Intel does.

Hopefully Intel will right this wrong, though in the current competitive climate we can only hope.

AMD Announces 7nm EPYC CPUs and Radeon Instinct GPUs; Intel 10nm Still Nowhere in Sight

AMD’s 7nm Vega 20 GPU die. Image credit: Anandtech

On November 6th, AMD held an event dubbed “Next Horizon,” during which they formally announced the next generation of EPYC “Rome” high-performance CPUs and Radeon Instinct machine learning/AI GPUs for the data center. These chips are manufactured on TSMC’s bleeding edge 7 nanometer fabrication process, said to deliver 2x the density and a 50% reduction in power consumption versus the currently used 14nm LPP node from GlobalFoundries. The day prior, Intel made somewhat of an attempt to upstage AMD, announcing its Cascade Lake-AP server CPUs, still manufactured on 14nm. Based on the specifications of the upcoming chips, AMD appears poised to take significant share from Intel in the data center market next year.

First, let’s dive into the known specifications of the EPYC chips:

CPU AMD EPYC “Rome” Intel Xeon “Cascade Lake-AP” AMD EPYC 7601 (“Naples”) Intel Xeon Platinum 8180M “Skylake-SP”
Node/uArch 7nm Zen 2 14nm++ Cascade Lake 14nm Zen 14nm+ Skylake-SP
Cores/Threads 64/128 48/96 32/64 28/56
Clock Speed
1.8GHz (ES) / ? ? 2.2GHz / 3.2GHz 2.5GHz / 3.8GHz
L2 Cache 32MB 24MB 16MB 28MB
L3 Cache 128MB 66MB? 64MB 38.5MB
L4 Cache Maybe?
Memory Controller Octa-Channel DDR4 (3200?)
Up to 4TB per socket
Dodeca-Channel DDR4-2667
Up to 3TB per socket
Octa-Channel DDR4-2667
Up to 2TB per socket
Hexa-Channel DDR4-2667
Up to 1.5TB per socket
I/O 128x PCI-E 4.0 96x PCI-E 3.0 128x PCI-E 3.0 48x PCI-E 3.0
Socket Socket SP3 (LGA 4094) (<2P) BGA 5908 (<2P) Socket SP3 (LGA 4094) (<2P) LGA 3647 (<8P)
Price $$$$ $$$$$$ $4200 $13000

First we immediately take note of the sheer size of this monstrosity: 64 cores, 128 threads, 160MB combined L2 + L3 cache. AMD achieved this in part by moving to an even more modular architecture than that used in the original EPYC, coupled with a new version of Infinity Fabric to reduce latency and increase bandwidth.

Image credit: Tom’s Hardware

The new EPYC incorporates 8 tiny CPU dies, each containing a complex of 8 cores, tied together through a massive I/O die which is still manufactured on the 14nm node. Things like I/O controllers don’t shrink down to smaller nodes very well, and the benefits of doing so are negligible. This allows AMD to significantly improve yields and cost of production, as well as improve greatly upon the issues caused by non-uniform memory access, high latencies between dies and multiple hops for data through Infinity Fabric.

The I/O die contains the PCI-E controller, memory controller, and likely an L4 cache (although this remains unconfirmed). This eliminates NUMA and non-uniform memory latency, ensuring that only one hop to the I/O die is necessary and allowing the chip to actually behave like a single-socket part. The L4 cache if implemented would be fully inclusive of the L3 (which is already inclusive of the L2), meaning that any data needed to be pulled from another die’s cache would already be present in the I/O die, improving dramatically on “Naples”‘ wildly varying cache latency.

“Rome” also is the first x86 CPU to implement the PCI-E 4.0 specification, doubling bandwidth for peripherals like graphics cards to 64GB/s bidirectional. It also boasts new Infinity Fabric Links, offering 200GB/s of bidirectional bandwidth to compatible Radeon Pro/Instinct GPUs as well as between CPUs in a dual-socket configuration. This puts it miles ahead of Intel’s current Xeon offerings, which only output 48 PCI-E 3.0 lanes per socket. However, unlike with EPYC where dual-socket configurations use half of each CPU’s PCI-E lanes for inter-socket communications (and thus doesn’t increase the total lane count), Intel’s PCI-E lane count is unaffected in multi-socket configurations, thus a system with two Xeons supports up to 96 PCI-E 3.0 lanes. This still falls far short of EPYC with less than half the total I/O bandwidth.

AMD didn’t reveal clock speeds or exact performance uplift, but a footnote in their press release suggested a 29% IPC improvement over “Naples”. Even if this is a best-case scenario and typical workloads only see half the improvement, this is nonetheless very impressive. They discussed numerous architectural improvements, such as an improved front-end and branch predictor, lower latencies and increasing the FPU width to 256 bits. This means that they’re tackling the key weaknesses of their previous lineup versus Intel’s CPUs, which mainly include workloads that are latency-sensitive or utilize 256-bit AVX.

At the event, they demonstrated one 64-core “Rome” CPU being benchmarked in C-Ray against two 28-core Xeon Platinum 8180M CPUs (the top of the line from Intel, costing $13000 each) in a dual-socket config. The EPYC machine finished the benchmark 7% quicker than the Xeons. Furthermore, AMD hinted that power consumption would stay the same with “Rome” (180W TDP), whereas the two Xeons have a combined TDP of 410W and also require a chipset which consumes about 20W. If this is even remotely indicative of typical performance, this chip will put Intel in the toughest competitive position it’s been in since 2005.

Intel’s response, which will probably be released at least a quarter or two later than “Rome” if their recent antics are anything to go by, is a 48-core Cascade Lake-AP CPU. While last year, Intel famously berated AMD for using 4 “glued together” dies in “Naples,” this CPU will utilize two of Intel’s 28-core dies on an MCM package, migrated to the more refined 14nm++ node. Note that 4 cores have been disabled on each die, likely due to (what will anyway be) excessive heat and power consumption. As each die supports 6 memory channels, the CPU will support 12 channels of DDR4 memory. No other concrete information was announced by Intel, leading this author to think the product is nowhere near being launched, similar to their 28-core desktop CPU announced 5 months ago which is still nowhere in sight.

Regardless of when it comes out, Intel’s CPU will likely have a hard time competing. It’s still manufactured on 14nm, using two gigantic 698mm^2 dies (“Rome” uses dies in the 70mm^2 range), and all expectations are set on a TDP of 300W or higher compared to 180W for EPYC. Moreover, based on available data, it looks like “Rome” will have equal, if not measurably better performance per clock and per core compared to Intel’s aging “-Lake” architecture. With nearly double the power consumption, 3/4 the cores, a way higher price tag, and less than half the I/O bandwidth, it’s hard to see the appeal of this part compared to EPYC. The main thing Intel has going for it right now is its reputation — it’s thoroughly established and entrenched in the data center, with a reliable track record and countless existing contracts, whereas AMD was absent from this market for the past few years. “Rome” might just be enough to make many large customers change platforms, though.

AMD made no announcements regarding desktop parts, but rest assured that Zen 2-based chips will be coming to Socket AM4 in 2019. Beyond that is the realm of rumor and speculation.

The Radeon Instinct chips are a lot less exciting, so I’m not going to talk about them in as much depth. There are two models: the MI60, featuring 64 compute units and a TDP of 300W; and the MI50, featuring 60 compute units and a TDP of 150W. The chip used is a die shrink of the current 14nm “Vega 10” silicon to 7nm, dubbed “Vega 20”, with enhancements such as dedicated INT4/INT8 hardware delivering 59/118TOPS respectively and a bump to a 1:2 FP64:FP32 ratio from Vega 10’s 1:16. Otherwise, the architecture is the same. The core config of 4096 SP, 256 TMU, and 64 ROP is retained, while the number of HBM stacks is doubled (running at 2000MHz effective) for a total of 16GB/32GB (MI50/MI60) and 1TB/s bandwidth. These cards are marketed as being for AI and machine learning, but they are arguably even better suited for scientific and HPC workloads thanks to the FP64 units. In this area they will compete with NVIDIA’s Tesla V100 PCI-E card, which offers similar theoretical performance on the 12nm node. For machine learning, NVIDIA’s Tesla T4 (also 12nm) should offer superior performance at a fraction of the power draw (75W vs 150W/300W for Radeon Instinct). These cards are not being marketed to gamers and for gaming workloads we would not expect them to exceed 1080 Ti performance. We are eagerly awaiting AMD’s “next gen” graphics architecture, coming ~2020-2021, as GCN (which is approaching its 8th birthday) is simply not exciting anymore and does not compete strongly. However, AMD will likely be competing on price.

But back to the exciting stuff, the CPUs. Is it over? Is Intel finished? How will 7nm Ryzen materialize? Let us know what you think in the comments.



AMD Radeon RX 590: Too Little, Too Late?

RX 480 reference card, courtesy AMD.

The rumor mill has yet to cease churning with word of AMD’s upcoming RX 590 graphics card, based on GlobalFoundries’ 12LP process. The RX 590 is said to utilize a respin of the aging Polaris chip, known as Polaris 30, shrunk down to the 12nm node from 14nm. This may give AMD another ~200MHz of headroom to work with, but is it enough to make a dent in NVIDIA’s share of the market?

For background, AMD initially launched the Polaris architecture with the RX 480, using the 14nm Polaris 10 GPU in 2016. An optimization, known as Polaris 20 and released as the RX 580, was released in 2017, providing slightly higher clock speed headroom on the 14nm process at the expense of power consumption. Polaris 30 marks the third refresh of the Polaris architecture for AMD, two years later, while NVIDIA has already moved on from Pascal to Turing. However, Turing is currently limited to the ultra-high end (>$500) market. As a result, the RX 590 will be going up against the same GTX 1060 that the RX 480 battled two years ago, and that the RX 580 is still in a dead heat against. The specifications of these three cards are not substantially different:

Graphics Card Radeon RX 480 Radeon RX 580 Radeon RX 590 (TBC)
GPU 14nm Polaris 10 XT 14nm Polaris 20 XT 12nm Polaris 30 XT
Core Config 2304 SP, 144 TMU, 32 ROP 2304 SP, 144 TMU, 32 ROP 2304 SP, 144 TMU, 32 ROP
Core Clock 1266MHz 1340MHz 1545MHz
Memory 8GB 256-bit GDDR5-8000MHz 8GB 256-bit GDDR5-8000MHz 8GB 256-bit GDDR5-8000MHz
Memory Bandwidth 256GB/s 256GB/s 256GB/s
Thermal Design Power 150W 185W 185W?
Launch Date 06/2016 04/2017 11/2018
Price $239 $229 ???

The rumored 15% clock bump, given linear scaling, would put the RX 590 decidedly ahead of the RX 480/580 and GTX 1060, but still closer to 1060 levels of performance than 1070 (much less 2070). But can we expect linear scaling?

The main issue I see with Polaris 30 is that, according to rumors, it’ll be using the same memory controller and the same 8Gbps GDDR5 as the previous Polaris cards. The problem is that Polaris is moreso limited by memory bandwidth than it is by raw shading, texturing, or rasterization performance. To a certain point, depending on game/workload, overclocking the memory is more beneficial than overclocking the core. AMD’s equally-performing card from the previous generation, the R9 390X (Hawaii), utilized a 512-bit bus with 6Gbps GDDR5, delivering 50% greater bandwidth than Polaris. More efficient compression algorithms (36% more, to be precise, not 50%) and other optimizations led to this bandwidth deficiency being negligible at the original 1266MHz stock clock, but how far can AMD push the envelope before it becomes pointless?

And moreover, it being two and a half years since Polaris launched, how did AMD lack the foresight to anticipate this refresh and the need for faster memory? 8Gbps may be the limit for stock GDDR5, but NVIDIA (or their board partners) utilized factory-overclocked 9Gbps GDDR5 for certain 1060 models. Given that AMD will presumably be launching a cut-down variant of this GPU, it would make sense for them to use binned chips to deliver higher bandwidth on the 590, then use the low bins for the cut-down card. However, according to current rumors, this will not be the case.

Another option AMD could have gone with would be to redesign, if nothing else (they didn’t redesign anything at all apparently), the memory controller. A 384-bit memory controller could provide for 384GB/s of bandwidth at 8Gbps, the same offered by a stock R9 390X. This seems a bit excessive for Polaris, so they could instead use 7Gbps GDDR5 and yield 336GB/s, which is more than enough, and offset the added cost and power consumption of the larger memory controller. Normally, this would also mean having to increase the rasterizer (ROP) count to 48, though if this were cost prohibitive AMD could have stuck with 32 on 384-bit as they did with Tahiti. A 384-bit, 48 ROP Polaris at 1600MHz though? Would that not be a 1070 competitor?

Practically coinciding with the launch of Polaris, NVIDIA launched Pascal with GDDR5X, which provided a bump to a ~10-11Gbps out-of-the-box data rate. Turing, launched this summer, uses GDDR6 running at 14Gbps. Across a 256-bit bus, 10Gbps delivers 320GB/s, and 14Gbps delivers 448GB/s — the same as the GTX 1080 and RTX 2080, respectively. If AMD could have simply redesigned the memory controller with the node shrink, instead of absolutely nothing at all, even 10Gbps GDDR5X would alleviate the bandwidth bottleneck, delivering a 25% increase in bandwidth versus 15% for the core clock. The main problem here is that memory is expensive, and neither of the newer memory technologies are being produced in particularly large quantities.

This is not to say that the RX 590 will be a particularly bad card for the price when it launches. It’s expected to perform perhaps 10% better than the GTX 1060 at approximately the same retail price. But the 1060 is a 2 year old card, consumes far less electricity, is about to be refreshed with GDDR5X memory itself, and is likely to be replaced by a 2060 in a few months. AMD shouldn’t be refreshing Polaris a second time to edge past it, they should have an all-new chip that decidedly beats it after all this time.

Ultimately, what it comes down to is that AMD is designing the RX 590 as cheaply as they possibly can. Their R&D budget is evidently minimal. From what we’ve seen so far, this card will not deliver a single change except moving to the 12nm node and taking the ~10-15% extra frequency that comes along with it. If it’s similar to the 12nm shrink they did for Zen, they won’t even increase the density of the design, they’ll just increase the space between die elements to improve heat dissipation and frequency potential. After two years, this is the best they can do, finally beating the 1060 when NVIDIA is already starting to roll out 2000-series graphics cards. Now that their GPU division has gotten an overhaul with the departure of Raja Koduri, it’s about time their GPU architecture gets one too — they need it, fast.

9 Kratom Fast Facts: What Is It, Why the DEA Wants it Banned, and Why They’re Wrong


Debates are being sparked across the country over what, if anything, the government should do in response to the recent surge in popularity of the psychoactive herb known as kratom. Hailed by its proponents as a miracle drug, kratom is a tropical evergreen tree in the coffee family that grows in the rainforests of Southeast Asia. In the West, it is typically sold in the form of finely ground leaf powder, which is then brewed into tea or made into capsules. A typical dose can cost less than 50 cents. Kratom contains a medley of psychoactive and medicinal alkaloids, the primary effects of which include pain relief, muscle relaxation, and mild paradoxical stimulating/sedating effects on the central nervous system. Users claim it to be the most effective solution available for pain relief, without the severe side effects of opioid painkillers or even those of over-the-counter drugs like acetaminophen. Additionally, since its mechanism of action in the brain is somewhat similar to that of opioids, it has been widely used with great success for the treatment of opioid withdrawal. Many believe it holds the key to solving the opioid crisis, freeing addicts from the withdrawal symptoms that keep them stuck and providing a far less risky alternative to prescription painkillers for those who need them.

Despite this, many government agencies have expressed concern about kratom as a drug of abuse and have moved to ban it. In 2016, the United States DEA announced its intention to place kratom in Schedule I of the Controlled Substances Act, the same class as heroin and LSD. After a massive public and scholarly outcry, the DEA rescinded its plans. However, since then, many state government bodies have moved to ban the drug, and it is currently illegal in Alabama, Arkansas, Wisconsin, Vermont, Rhode Island, and Indiana. Furthermore, the FDA under Commissioner Scott Gottlieb has been pushing the DEA to act on a federal level. This has been accompanied by misleading and sensationalist claims, such as when the Ohio Board of Pharmacy said that most users are injecting the herb (which is impossible), or when the FDA released a list of “kratom deaths” in which every single individual minus one had another easily identifiable cause of death listed (the kratom in their system was coincidental). The misinformation, fear-mongering, and brash, hasty reactions calling for a flat-out ban are eerily reminiscent of the 20th century “reefer madness”; ironic considering that the US is just finally emerging from that misstep now. Stay tuned for more in-depth write-ups on the subject, but here’s the quick run-down on why the feds are dead wrong:

  1. Kratom isn’t a gateway to drugs — it’s a gateway to sobriety. People don’t use kratom to get high; more often they’re using kratom to stop getting high. If someone is looking to abuse a drug, kratom is not a very good option. It can only be taken orally, is very unpleasant to consume in excessive amounts, and produces effects that are mild and take time to set in. The vast majority of people using kratom, whether they’re using it to treat pain or addiction, are using it in order to get away from opioids. With that taken into account, why on earth would you ban it, unless you want those people back on opioids? Banning kratom would worsen the quality of life of millions of Americans, prevent patients from receiving a potential breakthrough treatment, and lead countless people back into opioid addiction — likely to result in their deaths.
  2. Kratom has a low, but not zero, dependence liability. Unlike prescription opioids or anti-anxiety medications, which can addict a user in 4 days, one has to take kratom non-stop throughout the day for months in order to become dependent on it. Dependency is more mild than that of anti-depressants, antipsychotics, or muscle relaxers (which are not controlled substances), let alone opioids or benzodiazepines. In fact, the most common treatment for opioid addiction in the US today is Suboxone, which is itself an opioid far more addictive than kratom. Withdrawal symptoms from kratom typically only include a runny nose, insomnia, diarrhea, and soreness. Compare to the absolute horror of opioid withdrawal, or to alcohol and benzodiazepine withdrawal which can often cause seizures, delirium, and even death. Also consider that the vast majority of those who are dependent on kratom were previously addicted to something that is infinitely more dangerous, whether under the supervision of a doctor or not.
  3. Kratom doesn’t cause users to be impaired, detached from reality, fiendish, strung-out, etc. Kratom has paradoxical stimulant effects which are often compared to those of coffee, but without the jitters. Compare to other pain medications and opioid replacement drugs, which are heavily intoxicating. Nobody is strung out because of kratom. Nobody lost their job or their family because of kratom. If anything, it’s the other way around.
  4. It is practically impossible to overdose on kratom. It’s possible in theory, sure — just like overdosing on THC. In practice, it’s not going to happen. Taking too large a dose generally only results in dizziness and vomiting at the worst — compare to Tylenol, which kills thousands in the US each year. However, more research needs to be done into what interactions between kratom and other drugs may exist, as these may (or may not) in rare cases be dangerous.
  5. While the alkaloids in kratom do bind to the brain’s opioid receptors, the mechanism of action is very different from that of opioids. The FDA’s claims otherwise are based on a computer simulation and not actual studies (which contradict them) — in other words, junk science. Kratom also affects a number of other sites in the brain, such as serotonin and noradrenalin receptors, which results in a wide variety of effects including reduced inflammation, blood pressure, and anxiety. Opioids of abuse are selective full agonists of the mu-opioid receptors, which means they specifically seek out that receptor type and fully activate its signalling to other parts of the brain. Kratom, on the other hand, acts as a partial agonist, which results in a more mild effect. The only other partial mu-opioid agonist available on the market is buprenorphine (Suboxone). As a result of this, both buprenorphine and kratom are unable to cause fatal overdose, as there is a “ceiling” to the level of effects that can be achieved with increasing dosages.
  6. However, there is another major difference that makes kratom unique from buprenorphine/Suboxone and all opioids. The mu-opioid receptor has two different signalling pathways, which result in differing effects. The first one, the G-protein pathway, is responsible for most of the desired effects of opioids — mainly relief from pain, anxiety, and certain gastrointestinal disturbances, as well as slight euphoria. The beta-arrestin pathway, on the other hand, is responsible for the severity of the side effects of opioids — constipation, dizziness, sedation, cognitive impairment, an extremely powerful reinforcing effect, strong addiction with hellish withdrawal symptoms, and, most importantly, respiratory depression — the cause of death in opioid overdose. Opioid drugs, and even buprenorphine/Suboxone, activate both of these pathways indiscriminately. Kratom, on the other hand, only activates the G-protein pathway, and thus does not produce strong euphoric effects or result in severe addiction.
  7. The War on Drugs has been a colossal and deadly failure. Use of controlled substances has skyrocketed since the Controlled Substances Act took effect in 1971. Billions, if not trillions of dollars wasted, millions of people thrown in prison, and innumerable lives ruined, only for it to have the opposite of the intended effect. Crack epidemic, meth epidemic, opioid epidemic — it’s abundantly clear that this does not work! We should not be expanding the War on Drugs, but rather looking towards alternative ways to deal with the drug problem in this country — and kratom is one of those alternatives.
  8. If kratom gets banned, further research into its uses and risk profile will not be able to occur. The government’s argument is that there’s a lack of clinical trials proving that kratom is any safer than opioids in humans, or that it works to treat any condition. They fail to take into consideration that there is no research proving the opposite either. The absence of evidence is not the evidence of absence, and banning the substance would make it impossible for such evidence to ever arise. This is similar to the case of cannabis, where it took decades for its actual medicinal potential to be uncovered due to draconian laws. If these bureaucrats were unbiased, why would they not wait for the research to be done? Based on this, it seems they’re not actually interested in the research — they’re just using this as an argumentative tactic.
  9. Pharmaceutical companies are aware of kratom’s potential and are scared of it cutting into their profits. Under the new Commissioner, Scott Gottlieb, the FDA has been pushing strongly against kratom, advising the DEA to place it in Schedule I. Mr. Gottlieb was previously on the board of the Big Pharma giant GlaxoSmithKline, which acquired two patents for alkaloids derived from kratom. These patents clearly described its low abuse potential, favorable side effect profile, and medicinal applications. Now Mr. Gottlieb wants to claim it’s the opposite, and has even gone as far as to encourage Americans on Twitter who take kratom to switch to the highly-addictive prescription opioid Suboxone instead! These corporations make tens of billions of dollars per year off of people who are addicted to their opioid drugs, funnel massive amounts of money into lobbying the federal government, and have a revolving door between their board rooms and executive positions in bureaucratic agencies tasked with regulating their very industry. In light of all the evidence, there is no alternative explanation that makes sense as to why the government wants to ban kratom — their claims fly in the face of everything we know about the herb, as well as reason and common sense.

Millions of Americans, including this author, have had their lives transformed by kratom. If you’d like to help fight the government’s efforts to ban this gift from Mother Nature, please visit the website of the American Kratom Association and consider making a donation to support their research, lobbying, and awareness efforts. If you have a story regarding kratom or any thoughts you’d like to share, by all means do so in the comments below.