<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	
	>
<channel>
	<title>
	Comments for The SSD Review	</title>
	<atom:link href="https://www.thessdreview.com/comments/feed/" rel="self" type="application/rss+xml" />
	<link>https://www.thessdreview.com/</link>
	<description>The Worlds Dedicated SSD Education and Review Resource &#124;</description>
	<lastBuildDate>Fri, 01 May 2026 21:52:53 +0000</lastBuildDate>
	<sy:updatePeriod>
	hourly	</sy:updatePeriod>
	<sy:updateFrequency>
	1	</sy:updateFrequency>
	
	<item>
		<title>
		Comment on Crucial P310 2280 1TB SSD Review &#8211; This SSD Will Knock Your Socks Off! by DoC		</title>
		<link>https://www.thessdreview.com/our-reviews/nvme/crucial-p310-2280-1tb-ssd-review-this-ssd-will-knock-your-socks-off/#comment-124891</link>

		<dc:creator><![CDATA[DoC]]></dc:creator>
		<pubDate>Fri, 01 May 2026 21:52:53 +0000</pubDate>
		<guid isPermaLink="false">https://www.thessdreview.com/?p=107928#comment-124891</guid>

					<description><![CDATA[Thank you so much,i read all about this drive on internet telling me to avod it using as system disk. I brout it anyway and it is awesome for that price. Thank you for reviewing it .]]></description>
			<content:encoded><![CDATA[<p>Thank you so much,i read all about this drive on internet telling me to avod it using as system disk. I brout it anyway and it is awesome for that price. Thank you for reviewing it .</p>
]]></content:encoded>
		
			</item>
		<item>
		<title>
		Comment on Intel Core Ultra 200 Series Motherboards Are Not Achieving M.2 Slot Gen 5 14GB/s SSD Performance by jack hicks		</title>
		<link>https://www.thessdreview.com/our-reviews/nvme/intel-core-ultra-200-series-motherboards-are-not-achieving-m-2-slot-gen-5-ssd-14gb-s-performance/#comment-124661</link>

		<dc:creator><![CDATA[jack hicks]]></dc:creator>
		<pubDate>Tue, 28 Apr 2026 14:56:22 +0000</pubDate>
		<guid isPermaLink="false">https://www.thessdreview.com/?p=108969#comment-124661</guid>

					<description><![CDATA[This is such a frustrating find for anyone who just dropped big money on a Gen5 SSD and a Z890 board. The explanation about the I/O Extender Tile having higher latency due to the &quot;longer die-to-die data path&quot; is exactly the kind of technical detail that enthusiasts need to know before they start RMAing perfectly good drives. It&#039;s wild that even with a multi-die configuration, we&#039;re still running into these bottleneck issues.
I’ve seen similar latency headaches in my line of work with high-speed network devices. When you’re pushing 10GbE or even 25GbE through various network accessories and switches https://serverorbit.com/network-devices/network-accessories, even a tiny bit of extra latency in the controller can cause weird throughput drops that are a nightmare to track down. You think it&#039;s the cable or the port, but often it&#039;s just how the hardware is architected to handle data flow. It&#039;s a good reminder that &quot;theoretical maximum&quot; and &quot;real-world performance&quot; are rarely the same thing once you factor in the physical traces on a motherboard.
Do you think Intel can actually fix this with a microcode update, or is this &quot;die-to-die&quot; latency just a permanent physical limitation of the Ultra 200 architecture?]]></description>
			<content:encoded><![CDATA[<p>This is such a frustrating find for anyone who just dropped big money on a Gen5 SSD and a Z890 board. The explanation about the I/O Extender Tile having higher latency due to the &#8220;longer die-to-die data path&#8221; is exactly the kind of technical detail that enthusiasts need to know before they start RMAing perfectly good drives. It&#8217;s wild that even with a multi-die configuration, we&#8217;re still running into these bottleneck issues.<br />
I’ve seen similar latency headaches in my line of work with high-speed network devices. When you’re pushing 10GbE or even 25GbE through various network accessories and switches <a href="https://serverorbit.com/network-devices/network-accessories" rel="nofollow ugc">https://serverorbit.com/network-devices/network-accessories</a>, even a tiny bit of extra latency in the controller can cause weird throughput drops that are a nightmare to track down. You think it&#8217;s the cable or the port, but often it&#8217;s just how the hardware is architected to handle data flow. It&#8217;s a good reminder that &#8220;theoretical maximum&#8221; and &#8220;real-world performance&#8221; are rarely the same thing once you factor in the physical traces on a motherboard.<br />
Do you think Intel can actually fix this with a microcode update, or is this &#8220;die-to-die&#8221; latency just a permanent physical limitation of the Ultra 200 architecture?</p>
]]></content:encoded>
		
			</item>
		<item>
		<title>
		Comment on Intel Core Ultra 200 Series Motherboards Are Not Achieving M.2 Slot Gen 5 14GB/s SSD Performance by Dc y		</title>
		<link>https://www.thessdreview.com/our-reviews/nvme/intel-core-ultra-200-series-motherboards-are-not-achieving-m-2-slot-gen-5-ssd-14gb-s-performance/#comment-122824</link>

		<dc:creator><![CDATA[Dc y]]></dc:creator>
		<pubDate>Sat, 04 Apr 2026 14:31:41 +0000</pubDate>
		<guid isPermaLink="false">https://www.thessdreview.com/?p=108969#comment-122824</guid>

					<description><![CDATA[In reply to &lt;a href=&quot;https://www.thessdreview.com/our-reviews/nvme/intel-core-ultra-200-series-motherboards-are-not-achieving-m-2-slot-gen-5-ssd-14gb-s-performance/#comment-112742&quot;&gt;Les Tokar&lt;/a&gt;.

Nope, it didn&#039;t. Still same on tomahawk z890 with 270k plus.]]></description>
			<content:encoded><![CDATA[<p>In reply to <a href="https://www.thessdreview.com/our-reviews/nvme/intel-core-ultra-200-series-motherboards-are-not-achieving-m-2-slot-gen-5-ssd-14gb-s-performance/#comment-112742">Les Tokar</a>.</p>
<p>Nope, it didn&#8217;t. Still same on tomahawk z890 with 270k plus.</p>
]]></content:encoded>
		
			</item>
		<item>
		<title>
		Comment on Toshiba Announces PX04S Series of Enterprise 12Gb/s SAS SSDs &#8212; Up to 3.84TB Capacity by Les Tokar		</title>
		<link>https://www.thessdreview.com/daily-news/latest-buzz/toshiba-announces-px04s-series-of-enterprise-12gbs-sas-ssds-up-to-3-84tb-capacity/#comment-122785</link>

		<dc:creator><![CDATA[Les Tokar]]></dc:creator>
		<pubDate>Fri, 03 Apr 2026 14:15:52 +0000</pubDate>
		<guid isPermaLink="false">https://www.thessdreview.com/?p=88523#comment-122785</guid>

					<description><![CDATA[In reply to &lt;a href=&quot;https://www.thessdreview.com/daily-news/latest-buzz/toshiba-announces-px04s-series-of-enterprise-12gbs-sas-ssds-up-to-3-84tb-capacity/#comment-122770&quot;&gt;DarkServant&lt;/a&gt;.

Thank you for taking the time to write and I couldn&#039;t agree with you more. Feel free to jump in with similar educated comments whenever you get the urge!]]></description>
			<content:encoded><![CDATA[<p>In reply to <a href="https://www.thessdreview.com/daily-news/latest-buzz/toshiba-announces-px04s-series-of-enterprise-12gbs-sas-ssds-up-to-3-84tb-capacity/#comment-122770">DarkServant</a>.</p>
<p>Thank you for taking the time to write and I couldn&#8217;t agree with you more. Feel free to jump in with similar educated comments whenever you get the urge!</p>
]]></content:encoded>
		
			</item>
		<item>
		<title>
		Comment on Toshiba Announces PX04S Series of Enterprise 12Gb/s SAS SSDs &#8212; Up to 3.84TB Capacity by DarkServant		</title>
		<link>https://www.thessdreview.com/daily-news/latest-buzz/toshiba-announces-px04s-series-of-enterprise-12gbs-sas-ssds-up-to-3-84tb-capacity/#comment-122770</link>

		<dc:creator><![CDATA[DarkServant]]></dc:creator>
		<pubDate>Fri, 03 Apr 2026 03:47:50 +0000</pubDate>
		<guid isPermaLink="false">https://www.thessdreview.com/?p=88523#comment-122770</guid>

					<description><![CDATA[In reply to &lt;a href=&quot;https://www.thessdreview.com/daily-news/latest-buzz/toshiba-announces-px04s-series-of-enterprise-12gbs-sas-ssds-up-to-3-84tb-capacity/#comment-22779&quot;&gt;Bytales&lt;/a&gt;.

TEN Years Later:  122TB SSD&#039;s are a reality and soon double, but they are all QLC 3D-NAND based, and are mostly off limits to consumers, not only because of high pricing - availability is a big problem.
AI has changed the Landscape, and we are in a never before seen crisis, DRAM prices got up by 300% to 400% or sometimes more (for DDR5 RDIMM&#039;s). HBM... the &quot;VRAM&quot; for A.I. GPU&#039;s which eat up an extreme amount of wafers - they have up to 288GB of &quot;on-chip&quot; VRAM and cost easily 50&#039;000$. The few big DRAM producers have contracts for years and there is simply no more left for the small people.
15,36TB NVMe PCI 5.0 enterprise SSD&#039;s don&#039;t cost 500$... they cost you about 7000US$ (April 2026) or more.
There was a technology which could have changed the SSD market forever, it was 3DXpoint aka Intel Optane, latency was reduced to near 10µs (write and read) the first generation (P4800X) had 30DWPD and later 60DWPD, the second generation (P5800X) was PCIe 4.0 and had a whopping 100DWPD, sizes maxed out at 1,5TB for the first generation and 3,2TB for the second generation.
It was based on some sort of chalcogenide based phase-change memory with an ovonic threshold switch, which changed electrical resistance to save data. It was much less complex in terms of memor                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         y management, no DRAM buffer needed and is was faster than anything for many years. Only one core in the controller. Power consumption was high because of the memory chips power hunger.
But despite the high price of this new type of memory, Intel made a loss on every sale. The project was discontinued, and the entire memory division was sold to SK-Hynix, which created a new brand now called Solidigm.
SSD&#039;s with 1DWPD or less are standard, 3DWPD are the highest you can get. 7,68TB to 15,36TB is the new sweetspot with up to 2TB per package/chip. SSD-controllers with sometimes more than ten cores.
There was a time (July 2023) when NAND was really &quot;too&quot; cheap, due to an overproduction. But they will not do the same mistake again...]]></description>
			<content:encoded><![CDATA[<p>In reply to <a href="https://www.thessdreview.com/daily-news/latest-buzz/toshiba-announces-px04s-series-of-enterprise-12gbs-sas-ssds-up-to-3-84tb-capacity/#comment-22779">Bytales</a>.</p>
<p>TEN Years Later:  122TB SSD&#8217;s are a reality and soon double, but they are all QLC 3D-NAND based, and are mostly off limits to consumers, not only because of high pricing &#8211; availability is a big problem.<br />
AI has changed the Landscape, and we are in a never before seen crisis, DRAM prices got up by 300% to 400% or sometimes more (for DDR5 RDIMM&#8217;s). HBM&#8230; the &#8220;VRAM&#8221; for A.I. GPU&#8217;s which eat up an extreme amount of wafers &#8211; they have up to 288GB of &#8220;on-chip&#8221; VRAM and cost easily 50&#8217;000$. The few big DRAM producers have contracts for years and there is simply no more left for the small people.<br />
15,36TB NVMe PCI 5.0 enterprise SSD&#8217;s don&#8217;t cost 500$&#8230; they cost you about 7000US$ (April 2026) or more.<br />
There was a technology which could have changed the SSD market forever, it was 3DXpoint aka Intel Optane, latency was reduced to near 10µs (write and read) the first generation (P4800X) had 30DWPD and later 60DWPD, the second generation (P5800X) was PCIe 4.0 and had a whopping 100DWPD, sizes maxed out at 1,5TB for the first generation and 3,2TB for the second generation.<br />
It was based on some sort of chalcogenide based phase-change memory with an ovonic threshold switch, which changed electrical resistance to save data. It was much less complex in terms of memor                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         y management, no DRAM buffer needed and is was faster than anything for many years. Only one core in the controller. Power consumption was high because of the memory chips power hunger.<br />
But despite the high price of this new type of memory, Intel made a loss on every sale. The project was discontinued, and the entire memory division was sold to SK-Hynix, which created a new brand now called Solidigm.<br />
SSD&#8217;s with 1DWPD or less are standard, 3DWPD are the highest you can get. 7,68TB to 15,36TB is the new sweetspot with up to 2TB per package/chip. SSD-controllers with sometimes more than ten cores.<br />
There was a time (July 2023) when NAND was really &#8220;too&#8221; cheap, due to an overproduction. But they will not do the same mistake again&#8230;</p>
]]></content:encoded>
		
			</item>
	</channel>
</rss>
