Measure the video memory pressure that gives back unused pipelines without the free space inside the allocator's blocks.

pressure_release read the driver's figure (vk_mem_vram), which counts the free space inside VMA's blocks as used.
While those blocks held holes the use stayed over tex_vram_pressure however many pipelines went, so every 5 s another
batch of up to 256 prebuilt pipelines no draw had used yet was destroyed until none was left, and each later first
visit missed them again. The texture and mirror evictions moved to vk_mem_vram_live for the same reason; the
pipelines now use it too, and the comment and both log lines say that the free space inside the blocks is not
counted. The comment above vk_pipeline_sizes no longer says the table's entries are never taken out.

Not reproduced before the change (no available log reaches the threshold). vk_pipeline.c compiles with no new
warnings.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

runtime/vk_pipeline.c

@@ -5035,7 +5035,7 @@ static void fast_retire(Pipe *p) {
 // each one's compiled code in video memory until it is destroyed. In a 4K session with DLSS and frame generation the
 // device-local memory outside the allocator was 4.2 GB with 6,300 graphics pipelines alive, 3,589 of them not drawn
 // once in that run. On a card short of video memory those are the cheapest thing to give back: no draw has recorded
-// them. So while the device-local memory in use is over NX2NAVI_TEX_VRAM_PRESSURE percent of the budget (vk_mem_vram,
+// them. So while the device-local memory in use is over NX2NAVI_TEX_VRAM_PRESSURE percent of the budget (vk_mem_vram_live,
 // which NX2NAVI_VRAM_BUDGET_MB caps for a test), the autosave thread takes up to RELEASE_BATCH graphics pipelines that
 // no draw has looked up in this run out of the table every RELEASE_EVERY_MS, and destroys them with their entries once
 // the GPU is past the submission being recorded and one more (as fast_retire). A draw that needs one later misses it as
@@ -5045,6 +5045,9 @@ static void fast_retire(Pipe *p) {
 // memo (vk_exec_draw.c, the PM context) nor any caller holds their VkPipeInfo. Not taken: mesh pipelines (without
 // stage libraries their build is long), entries building or with a link queued, failed ones, and compute pipelines.
 // The next pass logs the device-local use then and now, so a log says what a batch gave back.
+// The use counts the free space inside the allocator's blocks as free, as the texture evictions do (vk_texture.c
+// sweep_pressure): the driver's figure (vk_mem_vram) counts it as used, so with free blocks it could stay over the line
+// while pipelines went, and every pass would give back another batch until no unused pipeline was left.
 // NX2NAVI_PIPE_RELEASE=0 keeps every pipeline for the run, as before 0.10.0.
 #define RELEASE_EVERY_MS 5000
 #define RELEASE_BATCH 256
@@ -5079,11 +5082,11 @@ static void pressure_release(uint64_t now) {
     if (!on || now - last < RELEASE_EVERY_MS) return;
     last = now;
     uint64_t used = 0, budget = 0;
-    int exact = 0;
-    if (!vk_mem_vram(&used, &budget, &exact) || !budget) return;
+    if (!vk_mem_vram_live(&used, &budget) || !budget) return;
     if (rel_used0) {
-        rt_log("vk_pipeline: device-local memory in use %llu MiB when the last batch of unused pipelines went, %llu MiB "
-               "now", (unsigned long long)rel_used0, (unsigned long long)(used >> 20));
+        rt_log("vk_pipeline: device-local memory in use (free space inside the allocator's blocks not counted) %llu MiB "
+               "when the last batch of unused pipelines went, %llu MiB now", (unsigned long long)rel_used0,
+               (unsigned long long)(used >> 20));
         rel_used0 = 0;
     }
     if (used * 100 < budget * (uint64_t)pct) return;
@@ -5103,9 +5106,10 @@ static void pressure_release(uint64_t now) {
     rel_total += (uint64_t)n, rel_batches++;
     rel_used0 = used >> 20;
     if (rel_batches <= 16 || !(rel_batches & (rel_batches - 1)))
-        rt_log("vk_pipeline: video memory in use %llu of %llu MiB (over %d%%): %d graphics pipelines no draw used in this "
-               "run given back, %llu in %llu batches so far; a draw that needs one builds it again from the pipeline "
-               "cache (NX2NAVI_PIPE_RELEASE)", (unsigned long long)(used >> 20), (unsigned long long)(budget >> 20), pct, n,
+        rt_log("vk_pipeline: video memory in use %llu of %llu MiB (over %d%%, free space inside the allocator's blocks not "
+               "counted): %d graphics pipelines no draw used in this run given back, %llu in %llu batches so far; a draw "
+               "that needs one builds it again from the pipeline cache (NX2NAVI_PIPE_RELEASE)",
+               (unsigned long long)(used >> 20), (unsigned long long)(budget >> 20), pct, n,
                (unsigned long long)rel_total, (unsigned long long)rel_batches);
 }
 
@@ -9250,8 +9254,9 @@ int vk_pipeline_selftest(void) {
     return ok ? 0 : 1;
 }
 
-// NX2NAVI_TRACE=perf "mem: sizes" (perf.c): the pipeline table's entries with their pipelines and fast links (never
-// evicted), layouts, stage libraries, the job queues and the pipeline-keys.bin set. A walk of the hash chains under P.lock
+// NX2NAVI_TRACE=perf "mem: sizes" (perf.c): the pipeline table's entries with their pipelines and fast links (only
+// pressure_release takes entries out, while video memory is short), layouts, stage libraries, the job queues and the
+// pipeline-keys.bin set. A walk of the hash chains under P.lock
 // held shared: the workers' inserts and a draw's memo miss wait for it (well under a millisecond).
 void vk_pipeline_sizes(PerfSizes *s) {
     uint64_t ent = 0, built = 0, fast = 0, libn = 0, libb = 0, lq = 0;